Editorial

Crunch time for tuberculosis control

Tuberculosis is a preventable and treatable disease that thrives amid poverty and weak health systems. For these reasons, the shortcomings and challenges in WHOs 13th annual report Global Tuberculosis Control 2009, released on March 24, make sombre reading. Tuberculosis incidence has declined since 2004, but only by less than 1% per year. Case detection is stagnant. Despite progress in several regions, the Stop TB Partnership target of a 50% reduction in 1990 prevalence and mortality by 2015 will not be met. More worrying is the synergy of tuberculosis and HIV/AIDS co-infection in sub-Saharan Africa, and the growth of multidrugresistant tuberculosis (MDR-TB) and extensively drugresistant strains (XDR-TB) in eastern Europe: two factors that complicate treatment and threaten to increase fatality rates. In 2007, there were 93 million incident reports of tuberculosis; half in Asia and a third in Africa. Overall, 13 million people were co-infected with HIV. 456 000 co-infected individuals died, making tuberculosis the commonest cause of death in people with HIV/ AIDS. Conversely, HIV/AIDS was responsible for almost a quarter of the 17 million deaths in people with tuberculosis. 500 000 people were thought to have MDR-TB and perhaps another 40 000 to have XDR-TB. India and China have the worlds largest burdens of incident tuberculosis, 20 million and 13 million people, respectively, and both have over 100 000 people with MDR-TB. The tragedysome might say follyis that established procedures for HIV/AIDS co-infection and MDR-TB have not been implemented widely. In 2004, WHO urged more collaboration between HIV and tuberculosis programmes, with routine testing for HIV in people with tuberculosis and for tuberculosis in people with HIV/AIDS. Testing accelerates diagnosis, improves treatment, and protects immunocompromised patients with HIV from tuberculosis. However, testing and treatment for the two infections often occur at separate sites. In 2007, the Global Plan targets for intensied case nding were missed worldwide. In Africa, the aim was to test 900 000 people with tuberculosis for HIV and 13 million people with HIV for tuberculosis. Only 500 000 and 300 000, respectively, were screenedand few of those diagnosed with co-infection received appropriate treatment.
www.thelancet.com Vol 373 April 4, 2009

MDR-TB is encouraged by poor case-detection, treatment with inappropriate drug regimens, and lack of clinical supervision. Almost half of treatment relapses in eastern Europe are from drug-resistant strains. Accurate diagnosis of MDR/XDR-TB requires drug sensitivity testing in a qualied laboratory. Only 2% of the estimated 500 000 people infected with drug-resistant strains were tested in 2007. And even when diagnosed with this more lethal form of tuberculosis, fewer than 3% received treatment recommended by international guidelines. Mechanisms to ensure best practice have failed at many levels in several countries because of lack of discipline, infrastructure, and resources. Clearly the changing nature of tuberculosis epidemiology demands a reassessment and scaling-up of control measures. To redene and redirect the actions necessary to combat tuberculosis, resolutions from the Stop TB Partners Forum will be presented in Beijing on April 13, when health ministers from the countries most aected by MDR/ XDR-TB will discuss strategies for tackling drug-resistant infection. To succeed they will need to build consensus, establish political will, and secure sustainable funding. The combination of MDR/XDR-TB, HIV/AIDS, weak health systems, and an extra 100 million people in poverty since the economic downturn in 2008, create ideal conditions for tuberculosis. Yet the same factors also provide targets to improve care through collaboration, strengthened health systems, and research. For example, the Article by Marcus Conde and colleagues in The Lancet today shows how a new drug might shorten treatment duration. Shorter treatments improve compliance and increase eciency. But more developments in diagnosis and treatment are needed. At a time when many US$ billions are spent on failing institutions, the underfunding by $16 billion a year for tuberculosis is shameful, particularly when each dollar spent on care generates $15 in productivity. Attitudes to tuberculosis must change among health professionals and the public. Laboratories and clinicians need to follow best practice in diagnosing, reporting, and managing the diseaseand they need to have the tools to do so. Additionally, eorts to control tuberculosis should engage communities to reduce stigma, support care, and develop local solutions. The meeting being held in China this week must be an inexion point in our collective response to tuberculosis. The Lancet

See Comment page 1148 See Articles page 1183

For Global Tuberculosis Control 2009 see http://www.who.int/ tb/publications/global_ report/2009/en/index.html

1145

Editorial

A right to a fair trial, a right to life

Since May 15, 2007, Binayak Sen, a distinguished Indian paediatrician and a tireless human rights activist has been imprisoned in a Raipur jail in the state of Chhattisgarh, India. He has been convicted of no crime but is being held under draconian state laws for his alleged association with the Naxalitesan outlawed Indian communist movement, deemed to be a threat to national security. To date, there is no proof of his involvement in extremist activities but he remains incarcerated for supplementary charges indenitely. Sen and his wife, Ilina, have devoted their entire working lives to improve the health and welfare of the Adivasis, a marginalised and poverty-stricken tribal population. Violent conict has prevailed in the region and Sens relentless exposure of the states human rights violations of this community are widely believed to be the real reason for his imprisonment: to set an example to others who would dare to expose state brutality and defend civil liberties. A troubling fallout of his incarceration is that much of his good work is slowly being eroded. His clinic, which provided essential health services, is on the verge of collapse, and many patients with both acute and chronic illnesses have gone untreated. The worldwide condemnation of his arrest and calls for his release continue to fall on deaf ears. Of grave concern now are reports that Sens health is deteriorating and that access to necessary medical care is being delayed. The right to life is a basic human right under the constitution of India. Every state functionary is obliged to protect the life of a detainee in custody and ensure proper medical treatment for him or her as and when required. It is outrageous that Sen has now been in prison for almost 2 years in a prolonged trial that keeps shifting charges which are unclear and possibly politically motivated. Faith in the Indian justice system needs to be restored. The Indian Government must intervene and make sure justice is done, so that Sen and his family can return to a normal life and resume serving the poorest communities in the state. The Lancet

For more on Binayak Sen and updates on his case see http://www.binayak.sen.net For more on the condemnation for Binayak Sens arrest see http://www.aidboston.org/ FreeBinayakSen/docs/binayak_ may_2007_textofpetition.pdf For more on the Constitution of India see http://indiacode.nic.in/ coiweb/welcome.html

An afternoon at UK Biobank
Before the age of 15, how many times did you suer sunburn that was painful for at least 2 days or caused blistering? How many children have you fathered? Do people say that you look younger than you are, older than you are, or about your age? How close to fact or ction are your answers? These are three examples of the 250 or so questions put to participants in UK Biobank, a research project aiming to enlist 500 000 people aged between 4069 years, and costing about 62 million so far. By analysing the answers; examining current and future medical histories; measuring baseline blood pressure, height, weight, grip strength, bone density, and lung function; and taking blood and urine for long-term storage and future tests, researchers may be able to work out why some people develop particular diseases while others do not. Perhaps that is what the eight or so healthy looking, white, middle-aged people who gathered on a damp afternoon in London hoped. Certainly, they are unlikely to benet directly. They relinquish all rights to their blood and urine samples, and give permission for access to their medical records at any time, even after death. No results are given, except for the baseline measurements taken at the visit, even if a life-threatening illness were discovered. All that, and participants face a lack of privacy when completing the computer-based questionnaire, and when answering more detailed questions about medical history. Participation in UK Biobank is voluntary, and can be withdrawn at any time. The consent process relies on reading the information leaet sent in the post and then ticking seven boxes on a computer screen on arrival. The assumption is that consent will be given. Sta are on hand, but the lack of privacy dissuades discussion. The success of UK Biobank depends on the right questions being asked, the continuing altruism of 500 000 volunteers, and for what the samples are tested. Lets hope that the data generate useful associations to predict and prevent disease in future generations. That would help transform an unpleasant afternoon into a useful one. The Lancet
www.thelancet.com Vol 373 April 4, 2009

Getty Images

For more on UK Biobank see http://www.ukbiobank.ac.uk For more on pilot phase of UK Biobank, and other biobank initiatives see Comment Lancet 2007; 369: 198082

1146

Comment

Bringing JUPITER down to earth

In The Lancet today, Paul Ridker and colleagues1 present a much awaited subanalysis from the JUPITER trial. In the initial report,2 these investigators tested and validated the hypothesis that asymptomatic individuals with normal LDL cholesterol concentrations but with evidence of inammation (increased C-reactive protein) would benet from treatment with a statin. Results of the new a-priori analyses show that asymptomatic individuals randomly allocated to rosuvastatin for just less than 2 years beneted particularly from this drug if low concentrations of both LDL cholesterol and C-reactive protein were achieved. These ndings add to evidence that cardiovascular disease has an inammatory component3 and suggest that inammation could become another target for primary prevention of cardiovascular disease. Although, in many guidelines, LDL cholesterol is recommended as the most important target for lipid-lowering treatment, Ridker and his team have shown previously that other variables of the lipid prolesuch as the cholesterol to HDL cholesterol ratioare, when combined with C-reactive protein, better predictors of coronary heart disease risk than is LDL cholesterol alone.4 Therefore, although this prespecied analysis on the basis of achieved LDL cholesterol is relevant, we should pay attention to the additional results of the report, which indicate that irrespective of the lipid endpoint used (including the apolipoprotein B to apolipoprotein AI ratio), a low concentration of C-reactive protein achieved with rosuvastatin treatment confers the best prognosis. What do we do with these ndings? Although JUPITER provides support to the lipid-inammation target as proof of concept, Ridker and colleagues now need to put their work into a clinical and public health perspective. For instance, many groups around the world have shown that a sedentary lifestyle, poor level of tness, abdominal obesity, smoking, insulin resistance, and metabolic syndrome are all predictive of raised concentrations of C-reactive protein.59 Although statins have enhanced greatly the ability to lower LDL cholesterol and thereby risk for cardiovascular disease,10 hopefully the ndings of JUPITER will spark further discussion on appropriate use of these drugs. The JUPITER results also indicate that we need to do a better job at global cardiovascular disease riskwww.thelancet.com Vol 373 April 4, 2009

assessment.11 Furthermore, despite the fact that regular exercise is a remarkable and cheap polypill, a comparison of the absolute reduction in risk for coronary heart disease that is gained after statin treatment versus that with a lifestyle-modication programme aimed at weight loss and improved tness is unlikely to happen in the near future. A low level of cardiorespiratory tness is a powerful predictor of cardiovascular disease independent of most studied risk factors.12 This issue is important with respect to prophylactic use of a powerful statin such as rosuvastatin in primary prevention. The gure shows a behaviour that has unfortunately become the norm rather than the exception. It is not meant to make trivial the contribution of statins to cardiovascular disease prevention. However, before we recommend lifetime treatment with rosuvastatin for millions of presumably asymptomatic individuals, we should remember that we do not have long-term safety data for this drug. Despite overwhelming evidence for the clinical benets of statins in terms of relative-risk reduction, absolute reduction in risk is the clinically relevant outcome. For example, even if raised concentrations of C-reactive protein in an individual increase the relative risk threefold, the patient could be at low absolute risk in the absence of other risk factors. From a pathophysiological standpoint, JUPITER provides key experimental data that inammation

Published Online March 29, 2009 DOI:10.1016/S01406736(09)60448-7 See Articles page 1175

1147

Hlne Trudel

Comment

is an important mediator of the clinical benets of rosuvastatin. However, to immediately translate these ndings into clinical practice without appropriate and careful discussion of their implications is not prudent. Hopefully, focus on the JUPITER trial will spur constructive and responsible dialogues to prioritise clinical and public health actions for primary prevention of cardiovascular disease. How can we delineate the proper population to treat? When should we use C-reactive protein in clinical practice? These issues will have to be examined carefully. Jean-Pierre Desprs
Centre de Recherche de lInstitut Universitaire de Cardiologie et de Pneumologie de Qubec, Qubec City, QC, Canada G1V 4G5 Jean-Pierre.Despres@criucpq.ulaval.ca
J-PD has been a speaker for Abbott Laboratories, AstraZeneca, Solvay Pharma, GlaxoSmithKline, and Pzer Canada; has received research funding from Innodia, Eli Lilly, GlaxoSmithKline, and Sano-Aventis; is on advisory boards for MSD, Novartis, and Sano-Aventis; and has received consulting fees from Innodia and Sano-Aventis. 1 Ridker PM, Danielson E, Fonseca FAH, et al, on behalf of the JUPITER Trial Study Group. Reduction in C-reactive protein and LDL cholesterol and cardiovascular event rates after initiation of rosuvastatin: a prospective study of the JUPITER trial. Lancet 2009; published online March 29, 2009. DOI:10.1016/S0140-6736(09)60447-5.

3 4

10

11 12

Ridker PM, Danielson E, Fonseca FAH, et al, for the JUPITER Study Group. Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein. N Engl J Med 2008; 359: 2195207. Libby P, Ridker PM, Maseri A. Inammation and atherosclerosis. Circulation 2002; 105: 113543. Ridker PM, Rifai N, Cook NR, Bradwin G, Buring JE. Non-HDL cholesterol, apolipoproteins A-I and B100, standard lipid measures, lipid ratios, and CRP as risk factors for cardiovascular disease in women. JAMA 2005; 294: 32633. Hak AE, Stehouwer CD, Bots ML, et al. Associations of C-reactive protein with measures of obesity, insulin resistance, and subclinical atherosclerosis in healthy, middle-aged women. Arterioscler Thromb Vasc Biol 1999; 19: 198691. Lemieux I, Pascot A, Prudhomme D, et al. Elevated C-reactive protein: another component of the atherothrombotic prole of abdominal obesity. Arterioscler Thromb Vasc Biol 2001; 21: 96167. Church TS, Barlow CE, Earnest CP, Kampert JB, Priest EL, Blair SN. Associations between cardiorespiratory tness and C-reactive protein in men. Arterioscler Thromb Vasc Biol 2002; 22: 186976. Ridker PM, Buring JE, Cook NR, Rifai N. C-reactive protein, the metabolic syndrome, and risk of incident cardiovascular events: an 8-year follow-up of 14 719 initially healthy American women. Circulation 2003; 107: 39197. Tracy RP, Psaty BM, Macy E, et al. Lifetime smoking exposure aects the association of C-reactive protein with cardiovascular disease risk factors and subclinical disease in healthy elderly subjects. Arterioscler Thromb Vasc Biol 1997; 17: 216776. Cholesterol Treatment Trialists (CTT) Collaborators. Ecacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90 056 participants in 14 randomised trials of statins. Lancet 2005; 366: 126778. Desprs J-P, Lemieux I. Abdominal obesity and metabolic syndrome. Nature 2006; 444: 88187. Sui X, LaMonte MJ, Blair SN. Cardiorespiratory tness as a predictor of nonfatal cardiovascular events in asymptomatic women and men. Am J Epidemiol 2007; 165: 141323.

Fourth-generation uoroquinolones in tuberculosis

See Editorial page 1145 See Articles page 1183

Drug development against Mycobacterium tuberculosis has been an overlooked area of tuberculosis research for some time. The rst breakthroughto obtain an active antituberculosis agentoccurred in the mid-1940s with the seemingly reluctant discovery of the activity of streptomycin1 and the more focused path towards the synthesis of para-aminosalicylic acid.2 The second breakthrough came when three drug companies simultaneously and independently applied to patent isonicotinic acid hydrazide (now known as isoniazid), only to learn that this drug had already been synthesised 50 years earlier.3 The third and perhaps last specic attempt at nding an antituberculosis agent came with the isolation of an antibiotic from the rifamycin class, published in 1959.4 Almost two decades passed before a regimen based on isoniazid plus rifampicin underwent rigorous assessment.5 Just 3 years after the identication of the rifamycins, nalidixic acid was synthesised.6 This agent was the starting point from which the later uoroquinolones

were developed. Fluoroquinolones were found to have activity against M tuberculosis and would become the rst real competitors of the two most powerful classes of antituberculosis agents that are now in routine use. Although nearly all uoroquinolones have antimycobacterial activity, the fourth generation, which includes gatioxacin and moxioxacin, have a particularly strong activity against M tuberculosis. In The Lancet today, Marcus Conde and colleagues7 report the eect on sputum-culture conversion at 8 weeks of treatment with isoniazid, rifampicin, and pyrazinamide given in combination with either moxioxacin or ethambutol (control). Ethambutol does not enhance, or only minimally enhances, the activity of the combination of isoniazid, rifampicin, and pyrazinamide against a fully susceptible strain. Indeed, so far, no drug has substantially enhanced the activity of this three-drug combination. The trials nding that culture conversion to negative occurred in 80% of patients in the moxioxacin
www.thelancet.com Vol 373 April 4, 2009

1148

Comment

group compared with 63% in the control group is, therefore, surprisingly large. Sceptics might argue that an improved 8-week culture-conversion rate suggests only superior bactericidal activity against the large bulk of rapidly metabolising organisms, whereas the crucial problem is the removal of persistent organisms that determine relapse frequency. The 8-week culture status is, nevertheless, a hallmark in assessing regimen ecacy, not least because it is a predictor of subsequent relapse.8 Conde and colleagues suggest that moxioxacin (and probably gatioxacin) has the potential to shorten treatment for tuberculosis, which is the ultimate goal given that relapse is only a small problem under routine programme conditions when the currently recommended 6-month regimen is administered properly.9 However, it remains to be seen whether fourth-generation uoroquinolones will allow tuberculosis treatment to be shortened. A point that deserves attention is the investigators discussion of the potential of this class in the treatment of multidrug-resistant tuberculosis. The biomedical literature has been inundated with XDR and similar attention-grabbing acronyms in countless commentaries and editorials. What seems often to have been lost in the discourse is the simple bottom-line shown by John Crofton10 in 1958: tuberculosis is easy to cure with one powerful bactericidal drug (then isoniazid) accompanied by a weak drug (then para-aminosalicylic acid), and supplemented by an injectable drug (then streptomycin), if the strain is susceptible to these drugs. Fourth-generation uoroquinolones probably have similar bactericidal activity to, and might exceed the sterilising activity of, isoniazid. Alternative injectable drugs to streptomycin frequently lack cross-resistance, and a weaker drug (or drug combination) with fewer and less severe toxic eects is often still available. In other words, multidrug-resistant tuberculosis (resistant to isoniazid plus rifampicin) that is not extensively drug resistant (additionally resistant to the uoroquinolones and injectable drugs other than streptomycin) should no longer be complex to treat. Yet international recommendations often seem very complex,11,12 suggesting questionable susceptibility testing to second-line drugs and the use of drugs with a poor record of adverse drug events, the latter forcing clinicians to change treatment or resulting in treatment default by the patient at a programmatically
www.thelancet.com Vol 373 April 4, 2009

unacceptable frequency.13 By contrast, attempts to cure extensively drug-resistant tuberculosis yield very poor results.14 Perhaps one lesson from this new trial is that there is great potential for treating uncomplicated multidrug-resistant tuberculosis with a simple standardised regimen containing a fourth-generation uoroquinolone. What is needed, and is perhaps in reach, is a regimen that is well tolerated, of reasonably short duration, without an unacceptably high frequency of adverse drug eects, and thus an eective treatment. Such a regimen will be deliverable at intermediate rather than specialised central levels in low-income countries. After all, it is these countries where the highest incidence of tuberculosis occurs and where complex recommendations that rely on sophisticated technology of doubtful value are unlikely to be successfully implemented. Hans L Rieder
International Union Against Tuberculosis and Lung Disease, 3038 Kirchlindach, Switzerland TBRieder@tbrieder.org
I declare that I have no conict of interest. 1 2 Comroe JH Jr. Pay dirt: the story of streptomycin. Part I: from Waksman to Waksman. Am Rev Respir Dis 1978; 117: 77381. Lehmann J. Twenty years afterward: historical notes on the discovery of the antituberculosis eect of para-aminosalicylic acid (PAS) and the rst clinical trials. Am Rev Respir Dis 1964; 90: 95366. Meyer H, Mally J. On hydrazine derivatives and pyridine carbonic acids. Monatshefte Chemie verwandte Teile anderer Wissenschaften 1912; 23: 393414 (in German). Sensi P, Margalith P, Timbal MT. Rifomycin, a new antibioticpreliminary report. Farmaco Ed Sci 1959; 14: 14647. Brouet G, Roussel G. Trial 6.9.12. Overall methodology and summary of results. Rev Fr Mal Respir 1977; 5 (suppl 1): 513 (in French). Lesher GY, Froelich EJ, Gruett MD, Bailey JH, Brundage RP. 1,8-naphthyridine derivatives. A new class of chemotherapeutic agents. J Med Pharm Chem 1962; 5: 106365. Conde MB, Efron A, Loredo C, et al. Moxioxacin versus ethambutol in the initial treatment of tuberculosis: a double-blind, randomised, controlled phase II trial. Lancet 2009; 373: 118389. Mitchison DA. Assessment of new sterilizing drugs for treating pulmonary tuberculosis by culture at 2 months. Am Rev Respir Dis 1993; 147: 106263. Dobler CC, Crawford ABH, Jes PJ, Gilbert GL, Marks GB. Recurrence of tuberculosis in a low-incidence setting. Eur Respir J 2009; 33: 16067. Crofton J. Sputum conversion and the metabolism of isoniazid. Am Rev Tuberc Pulm Dis 1958; 77: 86971. WHO. Guidelines for the programmatic management of drug-resistant tuberculosis. Geneva: World Health Organization, 2006. WHO/HTM/ TB/2006. http://whqlibdoc.who.int/publications/2006/ 9241546956_eng.pdf (accessed March 30, 2009). WHO. Guidelines for the programmatic management of drug-resistant tuberculosis: emergency update 2008. Sept 25, 2008. http://whqlibdoc.who. int/publications/2008/9789241547581_eng.pdf (accessed March 17, 2009). Nathanson E, Gupta R, Huamani P, et al. Adverse events in the treatment of multidrug-resistant tuberculosis: results from the DOTS-Plus initiative. Int J Tuberc Lung Dis 2004; 8: 138284. Keshavjee S, Gelmanova IY, Farmer PE, et al. Treatment of extensively drug-resistant tuberculosis in Tomsk, Russia: a retrospective cohort study. Lancet 2008; 372: 140309.

4 5 6

8 9 10 11

12

13

14

1149

Science Photo Library

Comment

PCI or CABG in coronary artery disease?

Published Online March 20, 2009 DOI:10.1016/S01406736(09)60574-2 See Articles page 1190

In The Lancet today, Mark Hlatky and colleagues1 report a pooled analysis of individual data from almost 8000 patients enrolled in ten randomised trials of percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG) over the past two decades. They conclude that, while at a median 6 years follow-up there was no overall dierence in survival, there was a signicant survival advantage with CABG in patients with diabetes (hazard ratio 070, 95% CI 056087) and in those aged 65 years or older (082, 070097). Furthermore, the combined endpoint of death or repeat revascularisation was reduced with CABG (10%) compared with PCI (25%; 041, 037045). Being probably the most denitive and authoritative analyses of the previous randomised trials, these conclusions are important and raise three important questions: are the ndings robust; are they consistent with previous reports; and are they generalisable to most patients undergoing PCI or CABG? The last question is particularly relevant to the recent publication of the 1-year interim-analysis of the landmark SYNTAX trial.2 First, however, it is necessary to consider two potentially important limitations of the new analyses. Most importantly, the randomised trials only enrolled around 510% of the eligible population, most of whom had single-vessel or double-vessel disease and normal left ventricular function,3 a group in whom it was

already well established that there was no prognostic benet with CABG.4 By largely excluding patients with a known survival benet from CABG (left-main or triple-vessel coronary artery disease, or both, and especially with impaired ventricular function4), the trials ignored the prognostic benet of surgery in more complex coronary artery disease. Nevertheless, inappropriate generalisation of trial results from highly selected populations to most patients with multivessel disease has been ubiquitous in the literature and has, at least partly, justied the explosive growth in PCI in developed countries. Similarly, reports from several large registries of a consistent survival benet of CABG over PCI in risk-matched patients with more complex coronary artery disease are often ignored.510 In fact, the severity of coronary artery disease in todays analyses is more similar to that recently reported in the COURAGE trial which showed no prognostic benet of PCI over optimum medical therapy,11 which implies that many of these trial patients could now be managed with medical therapy rather than any intervention. A second obvious limitation is that neither the PCI nor the CABG in these trials would be considered optimum by contemporary standards. PCI patients did not receive drug-eluting stents and only 83% of CABG patients received an internal mammary artery, the most important prognostic factor for long-term survival after CABG and a benet which persists long into the second decade of follow up.12,13 Furthermore, although use of bilateral internal mammary arteries can oer even greater prognostic benet,14 best evidence shows that, although drug-eluting stents reduce the incidence of restenosis compared with bare-metal stents, they do not improve survival or reduce the incidence of myocardial infarction.15,16 These observations are consistent with the hypothesis that, whilst bypass grafts to the mid-coronary vessel both treat the culprit lesion and also oer prophylaxis against new proximal disease, stents in the proximal coronary artery cannot protect against new disease. With these caveats and for the rst question above about robustness, Hlatky and colleagues, who are an eminent group of clinical and scientic academics in cardiovascular medicine (although without a surgeon among the 24 authors), argue persuasively that, by
www.thelancet.com Vol 373 April 4, 2009

Science Photo Library

1150

Comment

contrast with previous meta-analyses, pooling data from individual patients provides more precise estimates of the eect of CABG and PCI on survival both in the total population and in subgroups. For the second question about consistency, outcomes in todays study are broadly consistent with previous reports of a minor17 or no18 survival advantage of CABG over PCI, but with a far higher need for repeat intervention with PCI. Todays novel ndings are of signicantly better survival with CABG in older patients (65 years) and in patients with diabetes. Although the mechanism(s) of improved survival in these groups is not known, intuitively such mechanisms are most likely to reect more advanced coronary disease, particularly so in diabetic patients. The survival benet in patients with diabetes persisted even when the results of the BARI trial were excluded, but is consistent both with registry data19,20 and the recent BARI report that the survival benet for CABG persists at 10 years.21 On the other hand, Hlatky and colleagues observation that there was no signicant eect of the extent of coronary artery disease on the relative eectiveness of PCI and CABG on survival is counterintuitive and at odds not only with the ndings in older patients and those with diabetes in the new analyses but also with previous meta-analyses,4 registry data,510 and the SYNTAX trial.2 The third and most important question is how generalisable are these results to the population of patients with more severe coronary artery disease who require intervention? Despite the reservations above, new important evidence can be found in the interim analyses of SYNTAX in 1800 patients with left-main and/or three-vessel coronary artery disease who were randomised to PCI or CABG.2 The unique strength of SYNTAX was not only as an all-comer trial of patients with the most complex coronary artery disease but also the maintenance of a parallel registry of patients excluded from randomisation (1077 in the CABG group whose disease was too complex for PCI, and 198 in the PCI group considered to be at excessively high surgical risk). At 1 year (with nal analyses at 5 years), 12% of patients who had CABG and 18% of those who had PCI reached the primary composite endpoint of death, myocardial infarction, stroke, or repeat revascularisation. Although the dierence was largely driven by repeat revascularisation but with no signicant dierence in
www.thelancet.com Vol 373 April 4, 2009

mortality, PCI failed to reach the pretrial-specied criteria for non-inferiority, with the authors concluding that CABG remains the standard of care for patients with three-vessel or left main coronary artery disease (and by contrast with Hlatky and colleagues study, there was a greater benet with CABG in more severe disease). However, the 1-year result might greatly underestimate the survival benet of CABG, which registry data has consistently shown to accrue with time compared with PCI and usually reaches statistical dierence at 23 years.510 Furthermore, although all patients who had PCI received drug-eluting stents, fewer than 30% of those who had CABG beneted from the potential prognostic benet of bilateral internal mammary artery grafts.14 Finally, it is uncertain whether the higher incidence of stroke at 1 year with CABG (22% vs 06%) was largely procedural or a consequence of substantially inferior secondary prevention (including dual antiplatelet, statin, antihypertensive, and angiotensin-converting-enzyme inhibitor) than in the PCI group. So what can we conclude from the new study, especially in light of COURAGE and SYNTAX? For less severe coronary disease (mainly one-vessel or two-vessel disease and normal left ventricular function), there is little prognostic benet from any intervention over optimum medical therapy. In such patients who do require intervention, perhaps for symptomatic reasons, there is no obvious survival advantage for either PCI or CABG (at least in patients who are not diabetic), but there is a signicantly higher risk of repeat revascularisation with PCI. In patients with more severe coronary artery disease, and especially those with diabetes, CABG is superior in terms of survival and freedom from reintervention. However, SYNTAX also underlined that PCI is a good optionat least over the shorter termin patients who are ineligible for or who refuse CABG, and also the importance of rigorous secondary prevention in patients who have CABG. Finally, in view of the prognostic benet of surgery, a multidisciplinary team approach should be the standard of care when recommending interventions in more complex coronary artery disease, to ensure transparency, real patients choice, and genuine informed consent in the decision-making process. For elective patients, this approach will necessitate separation of angiography from the intervention to allow appropriate time to make a truly informed decision.
1151

Comment

David P Taggart
Nueld Department of Surgery, Oxford University, John Radclie Hospital, Oxford OX3 9DU, UK David.Taggart@orh.nhs.uk
I declare that I have no conict of interest. 1 Hlatky MA, Boothroyd DB, Bravata DM, et al. Coronary artery bypass surgery compared with percutaneous coronary interventions for multivessel disease: a collaborative analysis of individual patient data from ten randomised trials. Lancet 2009; published online March 20. DOI:10.1016/S0140-6736(09)60552-3. Serruys PW, Morice MC, Kappetein AP, et al, for the SYNTAX Investigators. Percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease. N Engl J Med 2009; 360: 96172. Taggart DP. Thomas B Ferguson Lecture: coronary artery bypass grafting is still the best treatment for multivessel and left main disease, but patients need to know. Ann Thorac Surg 2006; 82: 196675. Yusuf S, Zucker D, Peduzzi P, et al. Eect of coronary artery bypass graft surgery on survival: overview of 10-year results from randomised trials by the Coronary Artery Bypass Graft Surgery Trialists Collaboration. Lancet 1994; 344: 56370. Brener SJ, Lytle BW, Casserly IP, Schneider JP, Topol EJ, Lauer MS. Propensity analysis of long-term survival after surgical or percutaneous revascularization in patients with multivessel coronary artery disease and high-risk features. Circulation 2004; 109: 229095. Hannan EL, Racz MJ, Walford G, et al. Long-term outcomes of coronary-artery bypass grafting versus stent implantation. N Engl J Med 2005; 352: 217483. Malenka DJ, Leavitt BJ, Hearne MJ, et al, for the Northern New England Cardiovascular Disease Study Group. Comparing long-term survival of patients with multivessel coronary disease after CABG or PCI: analysis of BARI-like patients in northern New England. Circulation 2005; 112 (suppl 9): I37176. Smith PK, Cali RM, Tuttle RH, et al. Selection of surgical or percutaneous coronary intervention provides dierential longevity benet. Ann Thorac Surg 2006; 82: 142028. Bair TL, Muhlestein JB, May HT, et al. Surgical revascularization is associated with improved long-term outcomes compared with percutaneous stenting in most subgroups of patients with multivessel coronary artery disease: results from the Intermountain Heart Registry. Circulation 2007; 116 (suppl 11): I22631.

10

11

12

13

14

15

16

17

18

19

20

21

Hannan EL, Wu C, Walford G, et al. Drug-eluting stents vs. coronary-artery bypass grafting in multivessel coronary disease. N Engl J Med 2008; 358: 33141. Boden WE, ORourke RA, Teo KK, et al, for the COURAGE Trial Research Group. Optimal medical therapy with or without PCI for stable coronary disease. N Engl J Med 2007; 356: 150316. Loop FD, Lytle BW, Cosgrove DM, et al. Inuence of the internal-mammary-artery graft on 10-year survival and other cardiac events. N Engl J Med 1986; 314: 16. Lytle BW, Blackstone EH, Sabik JF, Houghtaling P, Loop FD, Cosgrove DM. The eect of bilateral internal thoracic artery grafting on survival during 20 postoperative years. Ann Thorac Surg 2004; 78: 200512. Taggart DP, DAmico R, Altman DG. Eect of arterial revascularisation on survival: a systematic review of studies comparing bilateral and single internal mammary arteries. Lancet 2001; 358: 87075. Kastrati A, Mehilli J, Pache J, et al. Analysis of 14 trials comparing sirolimus-eluting stents with bare-metal stents. N Engl J Med 2007; 35: 103039. Stettler C, Wandel S, Allemann S, et al. Outcomes associated with drug-eluting and bare-metal stents: a collaborative network meta-analysis. Lancet 2007; 370: 93748. Homan SN, TenBrook JA, Wolf MP, Pauker SG, Salem DN, Wong JB. A meta-analysis of randomized controlled trials comparing coronary artery bypass graft with percutaneous transluminal coronary angioplasty: one- to eight-year outcomes. J Am Coll Cardiol 2003; 41: 1293304. Bravata DM, Gienger AL, McDonald KM, et al. Systematic review: the comparative eectiveness of percutaneous coronary interventions and coronary artery bypass graft surgery. Ann Intern Med 2007; 147: 70316. Niles NW, McGrath PD, Malenka D, et al, for the Northern New England Cardiovascular Disease Study Group. Survival of patients with diabetes and multivessel coronary artery disease after surgical or percutaneous coronary revascularization: results of a large regional prospective study. Northern New England Cardiovascular Disease Study Group. J Am Coll Cardiol 2001; 37: 100815. Pell JP, Pell AC, Jerey RR, et al. Comparison of survival following coronary artery bypass grafting vs. percutaneous coronary intervention in diabetic and non-diabetic patients: retrospective cohort study of 6320 procedures. Diabet Med 2004; 21: 79092. BARI Investigators. The nal 10-year follow-up results from the BARI randomized trial. J Am Coll Cardiol 2007; 49: 160006.

Lipid lowering for primary prevention

Three large trials of rosuvastatin to prevent cardiovascular events have been completed.13 Two of these, CORONA and GISSI-HF,1,2 assessed 10 mg rosuvastatin daily. Substantial reductions in LDL cholesterol, a small increase in HDL cholesterol, and appreciable reductions in high-sensitivity C-reactive protein (hs-CRP) were reported. CORONA enrolled patients with ischaemic heart disease, whereas GISSI-HF included patients with ischaemic (40%), dilated (35%), and hypertensive (18%) causes of heart failure. Because such patients are probably at risk for future ischaemic vascular events, lowering of LDL cholesterol would be expected to reduce cardiovascular death, myocardial infarction, and stroke. Yet in neither trial was there a clear reduction in ischaemic vascular events or cardiovascular mortality (table).
1152

By contrast, in the JUPITER trial,3 17 802 apparently healthy people, with LDL cholesterol less than 34 mmol/L and CRP concentrations above 20 mg/L, received rosuvastatin 20 mg daily. LDL cholesterol decreased by 50% and CRP by 37%. Over 19 years, there were substantial and signicant reductions in ischaemic vascular events as well as total mortality, which were larger and more rapid than those in previous trials of rosuvastatin or other statins.4 How can we explain the apparently contradictory results of JUPITER compared with CORONA and GISSI-HF? In all three trials, CRP was raised, and substantial reductions in both LDL cholesterol and CRP occurred. The duration of the rst two trials was much longer than that of JUPITER, and because the benets of lipid lowering are enhanced by longer
www.thelancet.com Vol 373 April 4, 2009

Comment

follow-up, the larger and more rapid benet in JUPITER compared with previous statin trials is a surprise. For example, a 1 mmol/L lowering in LDL cholesterol reduces the risk of vascular events by 10% after 1 year,

25% after 23 years, and about 30% after 4 years.4 The 50% risk reduction in vascular events in JUPITER, after a 12 mmol/L drop in LDL cholesterol after 19 years, is unexpectedly large and rapid.
GISSI-HF (n=4574)2 JUPITER (n=17 802)3 Cholesterol Treatment Trialists (CTT) meta-analysis (n=90 056)4*

CORONA (n=5011)1 Background Age (years, mean) % women Major clinical diagnosis 73 24% Ischaemic heart failure (NYHA IIIV, LVEF <40%) Rosuvastatin (10) 27 35 45% 16 35 (1675) 37%

68 23% Heart failure (40% ischaemic) (NYHA IIIV) Rosuvastatin (10) 39 31 32% 10 27 (1260) 17%

66 38% Apparently healthy with high hs-CRP Rosuvastatin (20) 19 28 50% 12 42 (2871) 37%

62 24% Vascular disease, diabetes, primary prevention with increased risk factors* Simvastatin (2040); pravastatin (2040); lovastatin (2080); uvastatin (4080); atorvastatin (10) 47 38 29% 11 NA NA

Statin and dose (mg per day) Duration of follow-up (years) Baseline LDL cholesterol (mmol/L) % reduction in LDL cholesterol (placebo subtracted) Absolute dierence in LDL cholesterol (controlactive, mmol/L) Baseline hs-CRP (mg/L; median, IQR) % reduction in hs-CRP (placebo subtracted)

Major outcomes (number of events [event rate per 100 patient-years]) Composite of cardiovascular death, myocardial infarction, and stroke Control Active Hazard ratio (95% CI) Fatal and non-fatal myocardial infarction Control Active Hazard ratio (95% CI) Fatal and non-fatal stroke Control Active Hazard ratio (95% CI) Cardiovascular death Control Active Hazard ratio (95% CI) Non-cardiovascular mortality (includes deaths of unknown cause) Control Active Hazard ratio (95% CI) All-cause mortality Control Active Hazard ratio (95% CI) 759 (122) 728 (116) 095 (086105) 644 (72) 657 (74) 100 (090112) 247 (125) 198 (100) 080 (097097) 4354 (22) 3832 (19) 088 (084091)|| 166 (25) 147 (22) 087 (069109) 130 (15) 156 (18) 119 (094150) 204 (121) 163 (096) 080 (065098) 1801 (089) 1730 (085) 095 (090101) 593 (96) 581 (93) 097 (087109) 488 (55) 478 (54) 096 (085109) 43 (025) 35 (021) 080 (052127) 2553 (13) 2102 (10) 083 (079087) 115 (17) 103 (15) 088 (067116) 66 (07) 82 (09) 123 (089170) 64 (034) 33 (018) 052 (034079) 1617 (082) 1340 (068) 083 (078088) 149 (22) 124 (18) 082 (064104) 70 (08) 61 (07) 089 (063126) 68 (037) 31 (017) 046 (030070) 4420 (23) 3337 (17) 077 (074080) 732 (123) 692 (114) 092 (083102) 550 (62) 553 (62) 100 (089113) 157 (085) 83 (045) 053 (040069) 6033 (31) 4747 (24) 079 (076082)

NYHA=New York Heart Association class. LVEF=left-ventricular ejection fraction. NA=not applicable. IQR=interquartile range. *Includes four primary prevention trials (including one trial in people with diabetes), six secondary prevention trials, three trials involving mix of primary and secondary prevention, one trial in renal transplant recipients; 42 131 people with pre-existent cardiovascular disease and 47 925 people without pre-existent cardiovascular disease. In subset of 626 patients. Control=placebo for rosuvastatin trials and for 11 of 14 trials included in CTT meta-analysis; one trial in CTT meta-analysis compared statin with usual care, one used open-label statin vs no treatment, and one compared high-dose with low-dose statin therapy. Per 10 mmol/L reduction in LDL.Adjusted for hospitalisation for heart failure in previous year, previous pacemaker, sex, diabetes, pathological Q waves, and use of angiotensin-receptor blockers. ||A recent network meta-analysis6 which pooled 20 primary prevention statin trials (n=63 899) found that people treated with statin (lovastatin, pravastatin, uvastatin, atorvastatin) had relative risks of 093 (95% CI 087099, p=003) for all-cause mortality, 089 (081098, p=001) for cardiovascular death, 085 (077095, p=0004) for major vascular events, 077 (063095, p=001) for myocardial infarction, 088 (078100, p=005) for stroke, and 084 (066108, p=018) for revascularisations. These relative risks are similar to those observed in the CTT,4 and suggest that relative benets of statins are similar in primary and in secondary prevention.

Table: Major trials of rosuvastatin in context of other major statin trials

www.thelancet.com Vol 373 April 4, 2009

1153

Comment

Although JUPITER explicitly chose healthy people with raised CRP, most people in CORONA and GISSI-HF were likely to also have raised CRP. Moreover, in all three trials, rosuvastatin lowered CRP; yet in CORONA and GISSI-HF little overall benet was observed. A subgroup analysis from CORONA showed little benet of rosuvastatin in those with high concentrations of CRP. But a re-analysis5 suggested a 13% relative risk reduction in those with a CRP concentration of 20 mg/L or more, compared with no benet in those with CRP less than 20 mg/L. While subgroup results are intriguing, they suggest a far more modest eect than that observed in JUPITER and, being retrospective, require independent conrmation. While advanced myocardial disease in individuals in the heart failure trials might explain a lack of eect of rosuvastatin on mortality, specic ischaemic events, such myocardial infarction or stroke, would be expected to be reduced. In CORONA, there were fewer myocardial infarctions and strokes (table), but the benet was modest and not signicant. In GISSI-HF, the reduction of fatal and non-fatal myocardial infarction was also modest and there was no reduction in strokes (table). The benet of rosuvastatin for specic ischaemic events in the two trials seems to be less than what could be expected on the basis of secondary prevention trials, whereas the large eect in JUPITER was greater than that expected.6 The table compares rosuvastatin trials with a metaanalysis of previous primary and secondary prevention trials. This meta-analysis indicates that for about a 1 mmol/L reduction in LDL over 5 years, there is a 21% risk reduction in vascular events, and a 12% risk reduction in mortality. These estimates are similar in primary6 and secondary prevention trials,4 and are more modest than that in JUPITER. Furthermore, a recent trial of simvastatin and ezetemibe in mild-to-moderate aortic stenosis7 showed that a 50% reduction in LDL over 44 years led to a 17% decrease in vascular events and no reduction in mortality. Additionally, there was a reduction in non-cardiovascular deaths of about 20% in JUPITER (chiey due to fewer cancer deaths) that was as large as the reduction in cardiovascular deaths. This nding is not biologically plausible (2 years is too short to inuence cancer) and inconsistent with previous trials. Therefore, chance might have exaggerated the apparent benets in JUPITER (compounded by the trials early termination). In view of the results of all other statin trials, the real benets of lowering LDL by 12 mmol/L for about 2 years (as in
1154

JUPITER) is unlikely to be larger than a 2030% reduction in relative risk for ischaemic events, and a 10% reduction for total mortality. Both estimates are within the 95% CI of the estimates on specic events observed in JUPITER. What are the clinical implications of the recent trials? First, rosuvastatin seems safe in the medium term. Whether the excess in diabetes observed in JUPITER is a chance nding or real is unclear, because it has not been observed in other trials of statins. Indeed one trial with pravastatin reported lower rates of diabetes,8 but it would be prudent to systematically explore this point in all trials. Second, one would expect benets from rosuvastatin to be proportionate to the degree of lowering of LDL cholesterol and the duration of treatment, so the best estimate on specic ischaemic events might be the weighted average of all trials of rosuvastatin, when viewed in the context of all other previous statin trials.4,6 Third, because the role of CRP as a predictor of risk remains controversial,9 its role as a marker of preferential benets from statins should be evaluated in ongoing trials, and in a meta-analysis of all trials that have stored blood samples suitable for CRP measurement. Fourth, because in those without clinical cardiovascular disease the absolute benets are likely to be modest in the short term (the absolute risk of events in JUPITER was low, despite use of CRP as a marker of risk), much longer-term trials than those currently completed are needed to discover the full benets (which may increase over time) and safety of life-long use. Fifth, the eect of statins in several ethnic groups, such as Chinese people and south Asians, needs clarication. In view of the potentially large public-health and economic implications of widespread use of statins in apparently healthy individuals with average risk levels, conrmation of the long-term results of major lowering of LDL cholesterol (as can be safely achieved by rosuvastatin) is needed before potent statins are used widely in average-risk healthy people. Substantial risk reductions in cardiovascular disease are theoretically possible by combined lowering of blood pressure and LDL cholesterol in those with average levels of both risk factors and no apparent vascular disease, but this promising hypothesis needs assessment.10,11 Whether biomarkers such as CRP or N-terminal probrain natriuretic peptide7,12 will be better than simple clinical risk factors in identifying individuals who would benet from preventive strategies remains unclear. At present, tobacco avoidance, maintenance of optimum weight,
www.thelancet.com Vol 373 April 4, 2009

Comment

a prudent diet, and regular exercise should remain the foundations for prevention of cardiovascular disease in apparently healthy individuals with average risk factors. *Salim Yusuf, Eva Lonn, Jackie Bosch
Population Health Research Institute, McMaster University and Hamilton Health Sciences, Hamilton, ON, Canada L8L 2X2 yusufs@mcmaster.ca
We thank Paul Ridker, Aldo Maggioni, John McMurray, and Colin Baigent for providing additional unpublished analyses from the JUPITER and the GISSI-HF trials, and from the Cholesterol Treatment Trialists Collaboration; and Ann Raback for secretarial assistance. We are all investigators in the ongoing HOPE-3 trial that is evaluating combined lowering of blood pressure and lipids (with rosuvastatin) in primary prevention. This investigator-initiated trial is funded by the Canadian Institutes of Health Research through a peer-reviewed grant, and by AstraZeneca who manufacture rosuvastatin. SY has received research grants and honoraria from several companies that manufacture statins (AstraZeneca, Bristol Myers Squibb, and Merck), EL has received research grants and honoraria from several manufacturers of statins (AstraZeneca, Merck, Pzer), and JB has received grants from AstraZeneca. 1 2 Kjekshus J, Apatrei E, Barrios V, et al, for the CORONA group. Rosuvastatin in older patients with systolic heart failure. N Engl J Med 2007; 357: 224861. GISSI-HF Investigators. Eect of rosuvastatin in patients with chronic heart failure (the GISSI-HF trial): a randomised, double-blind, placebo-controlled trial. Lancet 2008; 372: 123139. Ridker PM, Danielson E, Fonseca FAH, et al, for the JUPITER Study Group. Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein. N Engl J Med 2008; 359: 219507.

10 11

12

Cholesterol Treatment Trialists (CTT) Collaboration. Ecacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90 056 participants in 14 randomised trials of statins. Lancet 2005; 366: 126778. McMurray J, Kjekshus J, Hjialmarson A, Wedel H, Dunselman P, Wongstein F. LDL independent reduction in cardiovascular morbidity and mortality with rosuvastatin in heart failure patients with a raised C-reactive protein: a retrospective analysis of the Controlled Rosuvastatin Multinational Trial in heart failure (CORONA). Circulation 2008; 118 (suppl 2): S711. Mills EJ, Rachlis B, Wu P, Devereaux PJ, Arora P, Perri D. Primary prevention of cardiovascular mortality and events with statin treatments: a network meta-analysis involving more than 65,000 patients. J Am Coll Cardiol 2008; 52: 176981. Rosseb AB, Pedersen TR, Boman K, et al, for the SEAS Investigators. Intensive lipid lowering with simvastatin and ezetimibe in aortic stenosis. N Engl J Med 2008; 359: 134356. Freeman DJ, Norrie J, Sattar N, et al. Pravastatin and the development of diabetes mellitus: evidence for a protective treatment eect in the West of Scotland Coronary Prevention Study. Circulation 2001; 103: 35762. Lloyd-Jones DM, Liu K, Tian L, Greenland P. Narrative review: assessment of C-reactive protein in risk prediction for cardiovascular disease. Ann Intern Med 2006; 145: 3542. Wald NJ, Law MR. A strategy to reduce cardiovascular disease by more than 80%. BMJ 2003; 326: 1419. Xavier D, Pais P, Sigamani A, Pogue J, Afzal R, Yusuf S, on behalf of The Indian Polycap Study (TIPS) Investigators. The need to test the theories behind the Polypill: rationale behind the Indian Polycap Study. Nature Clin Pract Cardiovasc Med 2009; 6: 9697. McKie PM, Rodeheer RJ, Cataliotti A, et al. Amino-terminal pro-B type natriuretic peptide and B type natriuretic peptide: biomarkers for mortality in a large community-based cohort free of heart failure. Hypertension 2006; 47: 87480.

The responsibilities of the World Medical Association President

Yoram Blachars accession to the Presidency of the World Medical Association1 comes with responsibilities. The World Medical Association was founded to work for the highest possible standards of ethical behaviour and care by physicians at all times. The Israeli Medical Association, of which he is also President, has on its website powerful position statements on health care during conict2 and on torture,3 which deserve strong support. Yet despite these statements, the Israeli Medical Association has been criticised for its reluctance fully to endorse international humanitarian codes.4 The Lancet has recently argued that it is the responsibility of medical professionals, and their professional bodies, explicitly to condemn unethical acts, even when such a challenge might prove unpopular.5 Blachars Presidency of the World Medical Association oers the opportunity to restore respect for the Israeli Medical Association from the global medical community, and creates opportunities for doctors to play a vital role in the search for peace. Israel has a right to act in pursuit of its security, but security might be used as a cover for many
www.thelancet.com Vol 373 April 4, 2009

authoritarian actions. Some actions of the Israeli General Security Services, under the umbrella of security, have superseded human rights, including the right to health care. Recent events in Gaza have led to widespread distress at the suering of civilians; international agencies have reported the Israeli Defense Force targeting medical stores and ambulances, and health workers have been killed. There were disturbing reports of the military refusing access to care for the injured, including one of children in a building found clinging to their dead mother 4 days after the house was shelled.6 Such acts are contrary to the principles stated in the Israeli Medical Associations paper, Assurance of medical and health services during the IsraelPalestine conict,2 as well as contravening international conventions. Yet the Israeli Medical Association has been disturbingly unforthcoming about these events. Blachar responded to my request for a statement of protest7 by claiming that Hamas was using medical facilities to store weapons and employing human shields.8 These reports are unveried and, according to

Published Online March 5, 2009 DOI:10.1016/S01406736(09)60373-1 See Comment Lancet 2009; 373: 78488

1155

Comment

international conventions, do not justify indiscriminate attacks on such targets. Just before the recent conict, I visited the occupied Palestinian territory. Health care in the West Bank, as for most aspects of daily life, has been completely undermined by the matrix of control exerted by the security forces. The separation wall and the checkpoints make travel dicult. We saw children in the East Jerusalem paediatric intensive care unit who had been transferred from Gaza without their parents being permitted to accompany them on security grounds, including newborn triplets.9 Teenage boys with type 1 diabetes are not allowed to cross the Israeli checkpoints to visit their specialist, because of ostensible security risks, even though their home and clinic are both on Palestinian land (Alem I, Ministry of Health, Nablus, occupied Palestinian territory; personal communication). The blockade of Gaza has exacerbated the need for patients to access health care in the West Bank and Israel. But in the rst 6 months of 2008, 34% of patients were refused permission on security grounds to leave Gaza for health care, compared with around 10% a year earlier, an increase in absolute numbers from 59 to 724 per month.10 It might be argued that, once explosives and weapons have been excluded, if it is still felt that patients present a security risk, they should be provided with military escorts. But perhaps in this setting, security has a dierent connotation. An August, 2008, report10 documents that the security grounds for refusing entry can be the refusal to provide information on Hamas

Mobile clinic in Gaza destroyed by air bombardment, January, 2009

suspects to the General Security Services. The Israeli Medical Association has not responded to this report. The Israeli Medical Association is to be congratulated for its clear statement on torture,3 ratifying the Tokyo Declaration of the World Medical Association, but there seems to be a disconnect between word and deed. The abuse of human rights by the Israeli Security and Prison Services has been documented. A May 2007, report, Ticking bombs: testimonies of torture victims in Israel,11 recounted the cases of nine Palestinian torture victims, with named medical personnel having been involved in their management. Again, the Israeli Medical Association has not responded to this report, and has not investigated the six doctors who are members. The Israeli Medical Association argues that its stance must reect its spectrum of membership. This might be why the position paper on health services during conict is in the international relations section of its English language website,2 but not in the Hebrew one (Weingarten M, Occupied Territories Project of Physicians for Human Rights-Israel; personal communication). It would seem unthinkable that Israeli doctors could condone attacks on health-care facilities or medical involvement in torture, so the Israeli Medical Association should be prepared to speak out, rather than appearing to neglect its humanitarian leadership role. Such an action might help divert calls for a boycott of the Association.4 Failing that, the World Medical Association President now needs to re-examine the situation from his new perspective, no longer hidebound by the constraints of the national association, to make a clear stand on each of these issues. Health care is a fundamental human right, and the medical profession worldwide should adhere to universal humanitarian codes. Doctors and medical associations in the region could build bridges to create conditions to help alleviate suering in the West Bank and Gaza. This could remove one of the grievances which drive people to violence. Without the tacit connivance of the medical profession, the General Security Services, the Defense Force, and the State would nd it dicult to continue with policies, including torture11 (even of children12) and denying patients access to health care.10 Such a move would require a bold step, but one which could secure Blachars place in history as a leader prepared to stand rmly for international humanitarian law, rather than one whose Association rationalises actions which are clearly unacceptable.
www.thelancet.com Vol 373 April 4, 2009

BTselem

1156

Comment

John S Yudkin
University College London, London WC1E 6BT, UK j.yudkin@blueyonder.co.uk
I thank Hadas Ziv and Miri Weingarten of Physicians for Human Rights-Israel for their help. I declare that I have no conict of interest. 1 World Medical Association. Health care systems must be protected from economic recession, says new WMA President. Oct 17, 2008. http://www. wma.net/e/press/2008_12.htm (accessed Feb 18, 2009). Israeli Medical Association. IMA position paper: assurance of medical and health services during the armed conict between Israelis and Palestinians. http://www.ima.org.il/en/CategoryIn.asp?cat1=18&cat2=20&id=20&tbl=tbl InternationalRelation2004&level=3# (accessed Feb 18, 2009). Israeli Medical Association. Prohibition of physician participation in interrogations and torture. December, 2007. http://www.ima.org.il/en/ CategoryIn.asp?show=Categories&id=317&tbl=tblCategoryabout (accessed Feb 18, 2009). Dyer O. Group of British doctors call for a boycott of the Israel Medical Association. BMJ 2007; 334: 871.

5 6 7 8 9 10

11

12

The Lancet. Violent conict: protecting the health of civilians. Lancet 2009; 373: 95. British Red Cross. Gaza is full-blown humanitarian crisis. Jan 7, 2009. http:// www.redcross.org.uk/news.asp?id=89867 (accessed Feb 18, 2009). Yudkin JS, Waterston T. Letter to Y Blachar. Dec 31, 2008 (available from JSY). Blachar Y. Response to reference 7. Jan 8, 2009 (available from JSY). Worth D, Metcalfe S, Boyd J, Worrall A, Canarutto P. Health and human rights in the Palestinian West Bank and Gaza. Lancet 2009; 373: 29596 Physicians for Human Rights-Israel. Holding health to ransom: GSS interrogation and extortion of Palestinian patients at the Erez Crossing. August, 2008. http://www.phr.org.il/phr/les/articlele_1217866249125. pdf (accessed Feb 18, 2009). Public Committee Against Torture in Israel. Ticking bombs: testimonies of torture victims in Israel. May, 2007. http://www.stoptorture.org.il/les/ 140%5B1%5D.pdf (accessed Feb 18, 2009). Defence for Children International, Palestine section. Israeli military continues to torture Palestinian children. June 26, 2008. http://www.dci-pal. org/english/display.cfm?DocId=798&CategoryId=1 (accessed Feb 18, 2009).

World Kidney Day: hypertension and chronic kidney disease

Hypertension is the most frequent complication of chronic kidney disease. Conversely, chronic kidney disease is a common and underappreciated medical cause of resistant hypertension. Adequate lowering of blood pressure to current guideline levels is arguably the most neglected part of the management of chronic kidney disease. This management decit prompted the International Society of Nephrology and the International Federation of Kidney Foundations to devote this years World Kidney Day (March 12, 2009) to Chronic kidney disease and hypertensiona marriage that should be prevented. The Day is a call to action to focus attention on the importance of chronic kidney disease, to raise awareness in the population, and to foster more aggressive intervention by doctors to reach blood pressure goals. Why is hypertension in chronic kidney disease relevant? The high and growing prevalence of earlystage chronic kidney disease as a contributor to raised cardiovascular risk has been appreciated only within the past decade. The increasing frequency of chronic kidney disease and end-stage kidney disease worldwide13 is, therefore, all the more alarming. In 2004, 19 million patients were on renal replacement therapy; by 2010, this gure is projected to increase by 40%.3,4 The most typical causes of chronic kidney disease are hypertension and diabetes. The kidney and hypertension have a fateful link.5 Reduced kidney function is a major cause of hypertension; conversely, hypertension is a leading factor in
www.thelancet.com Vol 373 April 4, 2009

initiation and promotion of progressive loss of renal function. Introduction of estimated glomerular ltration rate6 has enabled recognition of early stages of chronic kidney disease in many patients, particularly those with diabetes and hypertension. Evaluation of this measure as an index of kidney function should be complemented by assessment of urine for protein or albumin (preferred). Because chronic kidney disease is associated with a major increase in cardiovascular risk, achievement of blood pressure goals in patients with the disorder is important. Lowering of systolic blood pressure to targets less than 130 mm Hg both reduces cardiovascular risk and retards progression of chronic kidney disease in proteinuric cases.7 The rst aim of World Kidney Day 2009 is to focus on proper measurement of blood pressure and assessment of urine as part of the medical examination for high-risk patients (eg, those with metabolic syndrome, diabetes,

Published Online March 12, 2009 DOI:10.1016/S01406736(09)60355-X

1157

Comment

or a family history of chronic kidney disease). The second objective is to provide information to patients with chronic kidney disease about high blood pressure, self-measurement of blood pressure at home, and lifestyle modications such as reduction of sodium intake and weight loss. To implement these goals, three approaches are suggested. First, World Kidney Day will raise awareness in the medical community about the currently underappreciated prevalence of chronic kidney disease and the importance of achieving target blood-pressure values. Failure to achieve target blood pressure is shown by ndings of the US Kidney Early Evaluation Program:8 prevalence (86%), awareness (80%), and treatment (70%) in a screened cohort of patients with chronic kidney disease were high, but blood-pressure control rates were disappointingly low (13%). Second, the Day will alert patients with chronic kidney disease (and patients organisations) of the crucial importance of self-measurement and treatment of high blood pressure to reach target values. Finally, World Kidney Day will raise public awareness (and, particularly, that of health-care providers) of the importance and size of the hypertension link in chronic kidney disease, with particular emphasis on lifestyle modication. Specically, reduced sodium intake and weight loss will be highlighted for patients with early signs of chronic kidney disease. Underdiagnosis of chronic kidney disease and undertreatment of hypertension are worldwide issues. Awareness of chronic kidney disease is low, and even more disturbing is the scarcity of knowledge that hypertension and diabetes are major risk factors for chronic kidney disease. Even recognition of the presence of chronic kidney disease does not ensure adequate treatment.

Therefore misconceptions might be present on the part of the patient, the provider, or both. An important barrier to overcome these false impressions is education of doctors and patients, exemplied by the Kidney Early Evaluation Program of the National Kidney Foundation in the USA8 and the Commission for the Global Advancement of Nephrology research and prevention programme of the International Society of Nephrology that addresses kidney health globally.9 *E Ritz, G Bakris, on behalf of the World Kidney Day Organising Committee
Department of Internal Medicine, Nierenzentrum, D69100 Heidelberg, Germany (ER), and University of Chicago Medical Center, Chicago, IL, USA (GB) prof.e.ritz@t-online.de
The World Kidney Day Organising Committee is: William G Couser (USA), John Feehally (UK), Bernardo Rodrguez-Iturbe (Venezuela), Miguel C Riella (Brazil), Philip K T Li (China), Georgi Abraham (India), Joel D Kopple (USA), Allan J Collins (USA), Paul Beerkens (Netherlands), and Anne Wilson (Australia). We declare that we have no conict of interest. 1 2 3 4 Atkins RC. The epidemiology of chronic kidney disease. Kidney Int Suppl 2005; 94: S1418. Alebiosu CO, Ayodele OE. The global burden of chronic kidney disease and the way forward. Ethn Dis 2005; 15: 41823. Coresh J, Selvin E, Stevens LA, et al. Prevalence of chronic kidney disease in the United States. JAMA 2007; 298: 203847. Lea JP, McClellan WM, Melcher C, Gladstone E, Hostetter T. CKD risk factors reported by primary care physicians: do guidelines make a dierence? Am J Kidney Dis 2006; 47: 7277. Klahr S. The kidney in hypertension: villain and victim. N Engl J Med 1989; 320: 73133. Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function: measured and estimated glomerular ltration rate. N Engl J Med 2006; 354: 247383. Khosla N, Bakris G. Lessons learned from recent hypertension trials about kidney disease. Clin J Am Soc Nephrol 2006; 1: 22935. Rao MV, Qiu Y, Wang C, Bakris G. Hypertension and CKD: Kidney Early Evaluation Program (KEEP) and National Health and Nutrition Examination Survey (NHANES), 19992004. Am J Kidney Dis 2008; 51 (suppl 2): S3037. Perico N, Plata R, Anabaya A, et al. Strategies for national health care systems in emerging countries: the case of screening and prevention of renal disease progression in Bolivia. Kidney Int Suppl 2005; 97: S8794.

5 6

7 8

1158

www.thelancet.com Vol 373 April 4, 2009

World Report

Africa sees obstacles to guinea worm disease eradication

Experts are stepping up their eorts to eradicate guinea worm disease from the world. But the nal push will not be easy in the six African countries with remaining cases. Wairagala Wakabi reports.
Africa is edging closer to eradicating guinea worm disease, but insecurity in some endemic areas and fears of cross-border transmission are causing concern among those spearheading the eradication campaign. Last year saw a 51% reduction in the number of cases reported worldwide to 4647, and now it is hoped that guinea worm will be the next disease after smallpox to be eradicated. According to the Carter Center, which spearheaded the initiative since 1986, in conjunction with WHO and US Centers for Disease Control and Prevention, the number of cases worldwide fell from 35 million in 20 countries 22 years ago, and now only aects six African countries Sudan, Ghana, Mali, Ethiopia, Nigeria, and Niger. Philip Downs, a Carter Center expert, says that the character of the disease makes it easy to eradicate. It is easy to diagnose, and while there is no vaccine, it can be relatively easy to develop interventions, through the use of water lters and teaching those with the worms not to walk into water. It is also easy to apply chemicals to water to kill the larvae. Fresh funding of US$55 million announced last December by the Bill & Melinda Gates Foundation, the UK Department for International Development, and the Carter Center to support the eradication campaign is expected to drive the nal push towards eradication. While annoucing the funding, Carter Center founder, former US President Jimmy Carter, said: The reduction of guinea worm cases by more than 99% proves that when people work together, great positive change is possible. Yet, observers and campaign ocials alike continue to see obstacles to the eradication initiative as it reaches a decisive stage. Instability remains a
www.thelancet.com Vol 373 April 4, 2009

key challenge. In north-eastern Mali, where the disease is endemic, there has been a surge in violence that made interventions harder. Transmission in that region begins in May, and it is feared if the security situation remains

...it is hoped that guinea worm will be the next disease after smallpox to be eradicated.
bad then it will fuel transmission. In Sudan, remote areas and those which have in the past had insecurity in the south of the country have disproportionately higher infestation rates. In the secure north, the disease has been eradicated. Sudan currently has the biggest number of cases worldwide at about 3600, although the caseload fell by 39% last year. Emmanuel Otaala, Ugandas outgoing health minister, says there is need for cross-border surveillance, pointing to the possibility for countries such as Sudan to export the disease to neighbouring states. Burkina Faso and Togo have recently imported cases from Ghana, a secure country but one with 500 cases. We need to deal with imported cases by inaugurating guinea worm surveillance systems. We have been helping to set up basic infrastructure for disease surveillance, but more resources are needed from national governments, said Downs. The surveillance systems need to be sensitive enough so that if there are imported cases, they are able to detect them and prevent recontamination of water sources. Ghanas cases also have eradication ocials concerned. Most cases in this country are in the northern region, aecting large communities with populations of up to 20 000 people that live along major roads. You would

imagine that such big populations have safe water. But boreholes and wells break down or dry up, so people turn to water collected in dams, explained an ocial. He added that it is harder to set up surveillance systems in such semi-urban settings, although Ghana is beginning to establish surveillance systems in aected areas, and provide safe water sources. Since 1986, six previously endemic countries have been certied by WHO as disease free, and eight others that have stopped transmission are awaiting certication. It is expected that during 2009, only Ghana, Mali, and Sudan will still be transmitting the disease. Experts stress that the last cases of an eradication campaign are the most dicult and expensive to eliminate. Although infected cases drop, surveillance of countries, including the smallest communities in the most remote areas, needs to be intensied to prevent outbreaks and setbacks, they say. And, in the case of guinea worm disease, which has a 1-year incubation period, there is a very high cost of maintaining a broad and sensitive monitoring system and providing a rapid response when necessary.

Wairagala Wakabi

A Red Cross Society volunteer teaches children in Ghana about guinea worm disease

1159

The Carter Center/E Staub

World Report

Economic crisis highlights mental health issues in India

Charities and psychiatrists in India are reporting an increase in calls and visits from people with mental health problems amidst the nancial crisis. Patralekha Chatterjee reports from Delhi.
The global nancial crisis might have dimmed the lustre of Indias economic growth story but it has also had one positive outcome: the countrys mental health issues have been brought into the spotlight. Suicides generally increase during times of economic troubles, experts say, and Asians may be particularly susceptible as the region has among the worlds highest suicide rates, a recent Reuters report pointed out. The suicide rate in South Korea, for example, nearly doubled during the Asian nancial crisis 10 years ago, with experts blaming the increase on stress caused by job and income losses. Debt and distress has also driven tens of thousands of Indian farmers to commit suicide in the past two decades. India is particularly vulnerable because economic success in the past 15 years had raised expectations enormously, and many are having severe problems accepting their loss of income and status. Several recent suicides resulting from job losses have highlighted the increased need for mental health services in the country. In March this year, a 31-year-old unemployed software engineer was found dead in his north Delhi house. He had been in deep depression ever since he lost his job 4 months ago. In February, a 24-year-old banker in Chennai, southern India, hung herself, unable to cope with life after being made redundant. (NIMHANS), noted that: Morbidity on account of mental illness is set to overtake cardiovascular diseases as the single largest health risk in India by 2010. The preliminary ndings of a WHO-supported multicentre study on mental health in India indicates that about 10% of the population in India have mental health problems, a senior health ocial in Delhi told The Lancet. The study, whose ndings have not yet been made public, was coordinated by WHO and Indias Ministry of Health and Family Aairs, and done in 11 centres across India in collaboration with leading medical colleges. Behind these statistics is Indias epidemiological and demographic transition, which is leading to the emergence of non-communicable diseases as a major public health problem. Rapid urbanisation and industrialisation has triggered social upheaval, causing many people stress and anxiety. At the same time, traditional family support systems are disappearing while new forms of external emotional support are few and far in between. The challenges become apparent on a visit to a mental health outreach clinic in Chattarpur, an urban village in south Delhi. The clinica sparsely furnished room with stained walls and broken chairsis one of the two in the city, run by the government. It is housed inside a primary health care centre. A young psychiatrist is attending to a father and son duo. The son, a middle-aged man crouching on the oor, is the patient. Both are barely literate. Neither can coherently explain the symptoms, nor are they carrying earlier medical records. Persuading people who are in urgent need of treatment to come to the clinic is the rst hurdle, says Sumita Wadhwa, sta psychiatric nurse at the
www.thelancet.com Vol 373 April 4, 2009

...traditional family support systems are disappearing while new forms of external emotional support are few and far in between.
It is still too early to have denitive estimates of the eect of the nancial crisis on mental health but non-governmental organisations (NGOs) and psychiatrists are reporting a spurt in calls and visits. In the last 2 or 3 months, we have seen an almost 30% rise in the number of callers, says Lakshmi Vijaykumar, founder of the Sneha Suicide Prevention Centre based in Chennaia voluntary organisation that oers emotional support to people who are lonely, depressed, or suicidal. Achal Bhagat, a leading Delhi psychiatrist who is also aliated with Saarthakan NGO working in mental healthsays that the nancial crisis has worsened the situation for those who are already vulnerable. Many individuals with mental health disorders have stopped taking medicines or visiting psychiatrists because they are unable to pay for their treatment in the current situation, he points out. But the psychological problems of the urban middle class forced to confront a shrinking job market, layos, and pay cuts is only the tip of the iceberg. In 2008, a joint publication by Indias National Human Rights Commission and the National Institute of Mental Health and Neuro Sciences

Patralekha Chatterjee

A nurse and patient at a mental health outreach clinic in Chattarpur, south Delhi

1160

World Report

clinic. The second one is to persuade patients to accept prolonged use of medicines. Stigma surrounding mental illness and low awareness about available treatments push many people to faith healers whose prime advantage is their accessibility. The Chattarpur clinic oers a glimpse of the obstacles confronting Indias District Mental Health Programme (DMHP), a agship initiative of the central government. The DMHP aims to integrate basic mental health services into primary health care. But it is currently only being implemented in 125 of 625 districts in India. Part of the problem is human resources. In a country of over a billion people, there are only around 3500 psychiatrists. Not only is there an acute shortage of psychiatrists, but even this small number is not equitably distributed. 75% of the psychiatrists are located in urban areas and only 25% cover the rural pockets that account for 75% of the population, notes Bangalore Nanjundaiah Gangadhar, a professor of Psychiatry at NIMHANS. We have a shortage of 1600 psychiatrist social workers and 9000 psychiatrist nurses, Indias former health minister, Anbumani Ramadoss, lamented at a recent public function. However, as NIMHANS Gangadhar points out, the governments DMHP operational since the mid-1990s has led to a few thousand primary health care doctors and health workers being trained in the identication and treatment of selected mental disorders. But the DMHP, still requires a lot of ne tuning. In many states, there are not even enough people to train mental health professionals who can implement the District Mental Health Programme, says Jagdish Kaur, a senior Health Ministry ocial in Delhi. However, in Indias Eleventh Five Year Plan (20072012), mental health has moved up the priority list. In February this year, the federal government approved INR4740 million
www.thelancet.com Vol 373 April 4, 2009

(US$95 million) to tackle the human resource crisis derailing the National Mental Health Programme. The money will go towards establishing at least 11 centres of excellence of mental health and neurosciences by upgrading existing mental health institutes and strengthening other institutions. This move is expected to add 100 psychiatrists, 400 clinical psychologists, 400 psychiatric social workers, and 800 psychiatric nurses to Indias health workforce each year. The Medical Council of India is also working towards motivating medical college students to take up postgraduate courses in psychiatry to cover the shortage of manpower.

Delivering mental health care in India will require task shifting to community health workers who are trained and supervised...
Stop-gap measures are also being implemented. The health ministrys Jagdish Kaur says that the centre has asked state governments to train doctors in short-term courses in mental health so that they can supervise the DMHP. The psychiatry component in undergraduate medical curriculum is also likely to be increased. Additionally, the government plans to make all health professionals in a district undergo a week-long training in mental health. But Indias mental health burden is too huge to be tackled by mental health professionals alone. Delivering mental health care in India will require task shifting to community health workers who are trained and supervised, says Vikram Patel, professor of International Mental Health and Wellcome Trust senior clinical research fellow in Tropical Medicine, London School of Hygiene and Tropical Medicine. Patel is associated with Sangath, an NGO based in Goa, which is currently building an evidence base on the safety and benets of non-specialists delivering care.

WHOs Mental Health Atlas 2005 says that, as far as community care for mental health is concerned, India and south eastern Asia lag behind the rest of the world. But work is underway to close the gap. In Thiruporur, a rural part of Tamil Nadu, the Schizophrenia Research Foundation (SCARF)an NGO working for people with chronic mental illness based in Chennaihas used lay volunteers, selected with the help of village leaders, to take mental health care into the community. The volunteers, called community rehabilitation workers (CRWs), have been trained in the detection of mental disorders and to implement simple interventions, working closely with families of people with mental illness, and to make appropriate referrals, says Rangaswamy Thara, head of SCARF. This training was followed by periodic reinforcement sessions. Simultaneously, SCARF trained medical ocers and multi-purpose workers from the primary health care centre in basic mental health care. The selection of CRWs from the local population facilitated access to homes of people with mental illness. Patients identied in the community by the CRWs were treated by a psychiatrist and reviewed periodically. A similar procedure was followed in camps held in remote villages. Some simple interventions oered by the CRWs included emotional support to the client, educating families on mental illness, management of behaviour problems, ensuring drug compliance, training patients in self care, and initiating small businesses. Over a period of 5 years, more than 600 patients with mental illness were registered and oered treatment in the programme. Although many of the steps being taken to address mental health services in India will take time to show results, training members of the community in the identication of mental disorders will be vital, especially during these times of economic uncertainty.

Patralekha Chatterjee
1161

World Report

Remarks about women doctors cause fury in Czech Republic

A leading Czech doctor has caused controversy in his country by calling the feminisation of the health service a dangerous and unfavourable trend. Katka Krosnar reports from Prague.
The president of the Czech Doctors Chamber Milan Kubek is facing calls for his resignation after he claimed that the Czech health service was suering because of the work of women doctors. Almost 53% of the countrys doctors are women, according to ocial gures. Addressing a parliamentary health committee meeting to discuss the shortage of doctors in the Czech Republic, Kubek said the continuing feminisation of the health system was a dangerous and unfavourable trend aecting the health system. Female doctors have two roles: as well as their work they have to manage the family and caring for children, therefore the working commitment of women can never be the same as men, said Kubek. Kubek also said female doctors often avoided specialised medical positions where it was more difficult for them to juggle their careers with family life. In a presentation outlining the factors contributing to a shortage of doctors, Kubek highlighted low pay, the ageing of practising doctors, a lack of young doctors, and the exodus of doctors to work abroad as further issues putting the quality of care at risk. Doctors immediately attacked the remarks and the Czech Association of Clinic Specialists (SAS) has now called on Kubek to resign. Dr Kubek is head of an association which is compulsory for all doctors to be members of and which is there to protect ethics and morale. That is not a position which of the health committee and an oncologist, denied that women doctors worked less than men. A person in Dr Kubeks position should not have said what he did and he should apologise. I have been working as a doctor for over 25 years and have never felt that I did any less than my male colleagues, she said. Kubek later issued a statement addressing women doctors in which he oered an explanation of his commentsbut did not apologise for making them. In no way did I want to undermine or criticise your work, qualications and ability. My aim... was to highlight the fact that some specialties are not attractive for female doctors and one of the reasons is professional conditions...for example the level of pay which does not make it possible for some colleagues to organise childcare if they dont have help from their partner or parents, he said. Michal Sojka, spokesman for the Czech Doctors Chamber, said Kubek was merely trying to highlight the shortage of doctors in the country as Czech doctors leave to work abroad for better pay and conditions, and the dicult situation for women doctors. Female doctors tend to choose specialties where there is less out of hours work, such as paediatrics or general practice, because its very dicult for them to juggle family life, said Sojka. The fact is that it is usually women who take care of the family and there are worse conditions for female doctors in the Czech Republic than in other European Union countries. There is a lack of child care, no hospital nurseries, and no job sharing or option of working reduced hours when children are young, added Sojka.

It is nonsense to say that women cannot hold exactly the same positions as men and wrong to say that women are one of the causes of problems in the health-care system....
Dr Kubek should be holding, said Zorjan Jojko, chairman of SAS. It is nonsense to say that women cannot hold exactly the same positions as men and wrong to say that women are one of the causes of problems in the health-care system, added Jojko, a cardiologist based in Prague. In an open letter to Kubek, the SAS said it was shocked by Kubeks comments. Without any objective reason or proof you insulted female doctors, who represent more than half of the institution which you chair, the letter said. Member of Parliament (MP) Sona Markova, a member of the parliamentary health committee and chairwoman of a parliamentary commission on equal opportunities, said Kubek should apologise for his comments. If this was any country other than the Czech Republic Dr Kubek would be under strong pressure to resign, said Markova, although she does not share calls for his resignation, arguing that he has done a lot of good for doctors. She added: I could not believe what I was hearingI just shook my head in disbelief. MP Dagmar Molendova, a member of parliament who is also a member

The printed journal includes an image merely for illustration

CTK

Milan Kubek, president of the Czech Doctors Chamber

Katka Krosnar
www.thelancet.com Vol 373 April 4, 2009

1162

Perspectives

Book Meditations on pain

A few year ago, I addressed the World Congress on Pain in Sydney. In the cavernous convention centre, I asked the participants to think about a patient of theirs in pain and then to write down what they imagined that patient would say if they asked him or her tell me about your pain. Many participants told me afterwards how this exercise helped them connect to a singular life under pain, constricted by pain, made complex by pain. Although we are far away from any simple answers about pain and its treatment, those of us who treat pain will take all the help we can ndincluding from our own imaginationsto help those who suer it. Pain and its Transformations: The Interface of Biology and Culture brings us some answers to our questions about pain. It contemplates the ways that pain is perceived, interpreted, experienced, and endured. It visits the cultural, religious, and political uses of pain. An impressive cast of scholars, scientists, and writersincluding neurologist Howard Fields, cultural scholar Elaine Scarry, anaesthesiologist Cliord Woolf, and cultural psychiatrist Laurence Kirmayerwas assembled by psychiatrist/anthropologist Arthur Kleinman and his colleagues to gaze at pain, together, from directions so vastly distant that they are ordinarily mutually mute. Typically, transactions of colloquia make for quite deadly reading for all who were not there. Not this book; it took me months to read because it is a book that needs to be undergone. A homunculus for the subjects it treats, Pain and its Transformation cannot be grasped as a matter of cognition nor yet as a matter of aect or of belief. Like the topicsmemory, conviction, desiretreated by this book, it has to be given time to act on one. As one ruminates on pain, one realises how deep in the strata of life it lies. Pain underlies the distinctions
www.thelancet.com Vol 373 April 4, 2009

we make among biology, thought, action, meaning, and self. Enacting or mobilising all of these dimensions, pain makes us confront and forgo the unusual habit we have developed of sequestering one such level from the others. Contemplating pain puts one in a position to see the relations among body, narrative, and subject formation. This is the very problem being admired today by literary scholars, postmodern theorists, ethicists, and natural scientists, and pain is here oered as an instance in which the problem can be, if not worked out, at least beheld.

As one ruminates on pain, one realises how deep in the strata of life it lies. Pain underlies the distinctions we make among biology, thought, action, meaning, and self.
An exemplar that can help one consider other such fundamentals as pleasure, love, and the doing of good or evil, pain scrambles our categories, confounds our universitydepartmental mode of organising the world, and plunges us into the chaos and intricacy and resonance of ordinary life. Touchingly, it also brings us into congress with each other. It is not the academic brain or the intellectual apparatus that admits us to the position to contemplate pain. Pain and suering, instead, in the words of the 4th-century South Indian scholar Buddhaghosa, are intrinsic to the natural rise and fall of existence. As if all of us have become one breathing organism of existence, inhaling and exhaling quietly, we gather silently in the recognition of the reality of pain, its universality, its injustice, and its greed. Clues abound in this collection to sleuth about some fundamental questions concerning perception

and representation. We are gifted to insights from smart people who know things we all need to know in order to make sense of our world. Howard Fields raises the cover on a vault of neurobiological truthgems: The function of the brain has more in common with that of language than with that of other bodily organs, thereby shifting the ground of biological thinking from tissue to discourse. He goes on to state, so seemingly humbly, that the spatiotemporal pattern of activity of neurons in a network that produces an action (or perception) is a representation. This is the bridge that aestheticians and literary scholars have been seeking between thought and image. Fields solves a consequential problemfor me and my narrative medicine colleagues anywaybetween symptom and language, between, even more fundamentally, event and its account. This is the bridge between story and narrative. If the brain takes in that which it perceives in the form of representation, then there is no as-yet undiscovered step that must translate the perceived into the represented. It is the same thing. This book models the kind of courageous and generous collective action necessary for us today in dealing with pain. Not only does the collection expose the secrets of one discipline to others heretofore left out of its discourse. Pain and its Transformations also includes, with seeming egality, new conversants around the table. We learn from musicologist Elizabeth Tolbert, who writes about the ritual mourning of Finnish-Karelian communities, that the transformative function of the lament, indeed of all ritual genres, is ultimately about the desire for sociality and that such laments indeed have much to tell us about the socially grounded, emotionally

Pain and its Transformations: The Interface of Biology and Culture Sarah Coakley, Kay Kaufman Shelemay, eds. Harvard University Press, 2008. Pp 439. US$ 4995. ISBN 0-674-02456-7.

1163

Perspectives

motivated, and embodied nature of representation more generally. Kleinman hears Buddhist scholar Luis Gomezs observation that one type of suering he calls inherent pain is not the awareness that there is no pleasure or release from pain but rather the knowledge that pleasure never meets the expectations of desire and recognises in Gomezs radical openingup of the very ground of pain the incommensurability of the knowledges on display around this table. Added to this problem of incommensurability is, Kleinman suggests, the diculty of moving from representation to experience, and he wonders whether a nearing of these two categories might be found in techniques of experiential transformation. In a few pages of this book, then, we travel from the

meaning of cultural ritual to the interior experience of the most private pain or pleasure. Let me close with the suggestion that no one is left out of this conversation. Chinese historian and philosopher Tu Weiming states that the distinctiveness of being human lies in continuity with, rather than rupture from Heaven, Earth, and the myriad things. He then says:
The body so conceived is an attainment. We do not own our bodies. We become our bodies; as we learn to sit, stand, walk, run, and talk, we are empowered to express ourselves through our bodies. Although ontologically we are our bodies, in an existential sense, we must learn to become our bodies. The Confucian idea of the living body as the primary

datum and an irreducible reality is diametrically opposed to the Cartesian view that the body, contrasted with the thinking mind, is not essential for self-identity. For the Confucians, the body is the proper home for the heart-andmind. Furthermore, the heart-andmind manifests itself through the lived body and expresses its true nature by the experiences and feelings of the body. This is the reason that Mencius asserted that only the sage could bring the bodily form to fruition: Our body and complexion are given to us by Heaven. Only a sage can give his body complete fulllment.

This, I suggest, is commensurability.

Rita Charon
rac5@columbia.edu

In brief
Book Manga medic
Osamu Tezuka, considered by many to be the father of modern manga, the artful Japanese comics, left many works of his untranslated. The most famous of these, now being released in English editions, is a long-running series about shadowy surgeon, Black Jack, who metes out a strange form of justice with his scalpel. In a genre obsessed with the Byronic hero, Black Jack is archetypaldrawn with a billowy black cape, a face that resonates with intensity, and a shock of hair perpetually obscuring one eye but calling into focus the large scar on his left cheek. He is an unlicensed surgeon, operating beyond the bounds of polite society but so profoundly skilled that he is sought out by all those in despair. Like the eponymous maverick in the television series House, M.D., he is perennially doubted, his genius and morality covered up in an arrogant and unlikable exterior. Black Jacks take on society is one of anti-capitalist disgust, and he uses his exorbitant fees to hit the wealthy where they will hurt most. His sense of justice is keen, and he often deceives those who hire him for unsavoury purposes. In one episode, a wealthy businessmans lout of a son is involved in a car accident and the father arranges to have an innocent tailor framed and executed to provide donor body parts to reconstruct the son. Black Jack seems to perform the operation and collect his fee, but we later nd out that he let the son die and operated on the tailor to make him look like the dead son. Black Jack oers a combination of dramatic noir, wide-eyed comic book bathos, and a erce morality. The storylines deftly range from the parabolic to the ridiculous, while the graphics are sometimes slapdash and sometimes cinematic. Similarly, characters range from the intensely human protagonist with his many moods and subtle expressions to caricatured doll-like women and proto-simian businessmen with huge bellies and cartoon eyes. The most consistently impressive scenes, however, are in the operating theatre. Tezuka, who trained as a physician but never practised, seems to feel utterly comfortable with visualising such details as the ileocaecal junction in a child with intussusception, even if the story involves the famous surgeon operating from memory in pitch black during a hostage crisis. Tezuka is best known for his early, child-oriented series, Astro Boy, as well as Kimba the White Lion, but the Black Jack series is perhaps his nest work. In these English editions, we see the work of a fully realised artist in his prime and a protagonist who looks at moral obligations from multiple perspectives, but always askance.

Black Jack, vols 14 Osamu Tezuka. Translated by Camellia Nieh. Vertical, Inc, 2008, 2009. Pp 300. US$1695. ISBN 978-1-934287-27-9 (vol 1), ISBN 978-1-934287-28-6 (vol 2), ISBN 978-1-934287-41-5 (vol 3), ISBN 978-1-934287-43-9 (vol 4).

Noah Raizman
nraizman@gmail.com

1164

www.thelancet.com Vol 373 April 4, 2009

Perspectives

On reection The costs of care

The scene in Korle Bu Hospital, in the trac-choked city of Accra, Ghana, was one of the most shocking I have witnessed. In a side room o the third oor paediatric intensive care unit, I counted 27 newborn babies lying on four black mattresses on the oor. They were being held hostage by the hospital against payment of bills for their care. Their mothers, who had no hope of raising these sums, were allowed into the baby prison every 3 hours to feed them, before being banished again. Some babies had been detained for months. That was in 2005, but the image is etched on my memory as the clearest example, although admittedly extreme, of the harm done by charging patients for medical services that are both unavoidable and unpredictable. Medical charities such as Save the Children have lobbied vigorously for the removal of charges, which are seen as a deterrent to care, and governments of some developing countries have introduced free access as a way of boosting their nations health. Earlier this year that assumption was challenged by a fascinating study in Ghana of over 2000 households with children under 5 years old (PLoS Med 2009; 6: e1000007). Half were given free care and the remainder were made to pay, the normal practice in the region. The results showed, unexpectedly, that providing free care did not improve health, although it led to a slight increase in use of medical services. Anaemia, haemoglobin concentrations, parasite prevalence, and deaths were similar in both groups. What is the explanation? The authors suggest the increase in use of medical services by those with free care may have been too modest to make a dierence. Alternatively, the other costs of treatment in Ghana such as travel, and taking time o work, were more important than the small charge12 000 cedis (about US$1) for malaria treatment, for example. The issue of user fees is a fraught one. There is a strong equity argument for removing them and there may be other unmeasured eects. Clearly, there can be no justication for the ludicrous fees imposed on mothers in the intensive care unit at Korle Bu Hospital. But this nding is a reminder that we must not let our hearts rule our heads. In terms of maximising bang for bucks, abolishing user fees may not be the best way of increasing health gain. The other barriers to treatment need closer investigation rst.

Lifeline
Mary Moran practised emergency medicine for
13 years, followed by a diplomatic career, and 3 years with the Mdecins Sans Frontires policy team. In 2004, she set up an international health policy unit at the London School of Economics, and now directs the Health Policy Division of the George Institute for International Health in Sydney and London. What has been the greatest achievement of your career? Setting up my own health policy unit to try to nd answers to why neglected diseases were still so neglected, and how things could be improved. Who was your most inuential teacher, and why? My English Literature tutor at uni, who played period-style lute and was a Blake acionado. He expected our familiarity with the text to match his own, and refused to accept cant, clich, or pedestrian thinking: a baptism by re. How do you relax? Long hot baths with a glass of red wine and a book. Or swimming laps. What is your favourite lm, and why? I loved Pedro Almodovars All About My Mother. Like Hitchcock, Almodovars camera suggests but leaves room to paint in the emotions ourselves; and he is surprisingly tender and forgiving towards his characters. What is your greatest fear? Being buried alive, and anything bad happening to my son. What are you currently reading? Simon Schamas The American Future: A History; Mohsin Hamids The Reluctant Fundamentalist; and re-reading Michael Chabons The Adventures of Cavalier and Klay. What items do you always carry with you? Lucas Papaw Ointmenta lovely, sticky golden paste that works for everything: dry skin, burns, cuts, and scrapes. What is your worst habit? Eating muesli (standing up) for breakfast, lunch, and dinner when Im busy. What is the naughtiest thing you have ever done? If you were caught what, if anything, did you learn? I staged a fake suicide by jumping out of a third oor window (onto a ledge) to relieve the tedium of a highschool English lesson. I learnt two pieces of Realpolitik: punishment is likely to be less severe if you get good marks; and heartfelt protest doesnt necessarily deliver change. You can have dinner tonight with a famous person of your choice: who would it be? Im not good at keeping it to just one soGnter Grass for his dry asides and love of oal; Clint Eastwood, for his last lms; and Samuel Pepys, for his gargantuan appetite for life.

Jeremy Laurance
J.Laurance@independent.co.uk

www.thelancet.com Vol 373 April 4, 2009

1165

Perspectives

The art of medicine Banking on stories for healthier cognitive ageing

On a cold Christmas Eve in 1985, Bill Keane arrived in the darkened halls of the Neuropathology Laboratories at Johns Hopkins. He and a friend had risked an accidentstrewn highway to drive to Maryland from Pennsylvania; we needed what he carried as quickly as possible. Bills mother had just died after a 12-year struggle with dementia, and he was bringing us her brain. 2 years earlier, having heard of our research interests in Alzheimers disease, Bill had arranged to donate his mothers brain to our brain banka precious repository of generously donated tissues, to which eager neuroscientists from around the world would compete for access. For Bill, handing over his mothers brain produced palpable relief as he ended his journey of caregiving, and marked the beginning of a collaborative friendship and a new way of thinking about Alzheimers disease. Although at the time we were all hoping for successful biological treatments for Alzheimers disease, we sensed then that caring trumped curing. Bills commitment to his mother and her fellow suerers, and his subsequent professional contributions to making the culture in the long-term care eld less disease-oriented, were inspiring. They helped launch us on a professional trajectory during which we have come to recognise that peoples narrativesthe faith they put in science to nd a cure and the stories they tell to their aected loved ones about the nature of their ailments are ultimately far more important in understanding their journey from health to illness than the tissues we extract and analyse. Today, medicine is becoming diminished in almost every measure except its budget and its ego. Massive government investments in the USA and around the world are being made to collect biological specimens, as if the secrets of healing could be found in our component parts. Rather than the more complex brain tissues that we collected, biobanks are increasingly collecting uids such as blood to harvest cells that themselves contain the even smaller elements, the all-powerful genes, and, beyond them, the essence of life: DNA. It is necessary that we rise above this reductionism and remember the ancient truism that healers serve society (and life) by understanding their own stories and appreciating the stories of their patients. This is the true art of medicinea core component of the healing and holism that we have lost. The evangelists of genetics promise a future of personalised cures that will emerge from a mastery of the human genome. Such knowledge is exciting and of potential benet. Yet no matter how successful medicine is in expanding the frontiers of the genome, an individuals genes will be only one component of a broad, deep social and ecological matrix that aects health over our lives. To understand the multifactorial nature of health and disease in a community, it is essential to explore cultural contexts as well as individual subjectivities, and the best way to do that is through stories. Now is the time for a true personalised medicine rooted in the art of storytelling and the science of biology to assert itself in support of the quality of life of individual patients. Scientic advances in Alzheimers disease have led us down a narrow path on the journey to addressing the major societal challenge of age-related cognitive impairment. We need to break down the hyped-up language of exaggerated therapeutic breakthroughs. We now believe that Alzheimers disease is really a heterogeneous set of conditions intimately related to ageing processes that we should consider a family of disorders rather than a single condition. We have also come to recognise major overlaps among conditions previously thought to be discrete entitiesmost dementias feature mixed pathologies. An unrelenting, largely exclusive pursuit of better molecular understanding of brain ageing has led to greater confusionthe Alzheimers disease story has stalled out. A closer look at the 100-year old history of the disease tells us that even Alois Alzheimer himself did not believe the cases he observed in the early 20th century should be considered as separate disease entities. So how do we invigorate our scientic and clinical approaches to brain ageing? In Cleveland, we are turning to a new form of bankingthat of personal narratives
www.thelancet.com Vol 373 April 4, 2009

Private Collection / DACS / The Bridgeman Art Library

The printed journal includes an image merely for illustration

Autobiography (1945) by Marc Chagall

1166

Perspectives

from individuals in our community who are aected by brain ageing. The Cleveland Regional StoryBanka transdisciplinary planning eort of scholars in the health sciences, humanities, social sciences, and ethicsis a project designed to collect, store, and analyse individual stories related to health, illness, and disease to enrich knowledge of subjective health and illness experiences, and ultimately to improve individual and public health in our region and beyond. There are revealing conceptual parallels between biobanks and our StoryBank. Genes are a smaller component of cells that join together to form tissues. Tissues themselves are contained in organs which are contained in bodies that exist in complex ecosystems. At the highest (and most neglected) level of analysis, an evolutionary approach to medicine can show us how our genes are modied as a function of surrounding environmental conditions. Similarly, stories are composed of words ordered by syntax that form units of meaning. Stories are carriers of meanings, belonging to and forming the essence of individuals who, in turn, are members of local, regional, national, and international communities that originate, protect, and evolve those stories. Narratives change through complex political and cultural processes that at their heart depend on our awesome ability as a species to learn and gain wisdom from experience. So how do stories aect health? The narrative medicine movement has arisen in the past few decades heralding the belief that each individuals story is unique. Healing emerges from living out an integrated life narrative that promotes coherence and purpose. Doctor-authors such as Oliver Sacks, Arthur Kleinman, Rachel Remen, and Rita Charon promote the power of story to foster understanding and healing. They cite a need for doctors to have a narrative competence (as well as humility) when interacting with their patients. Intuition, clinical experience, and some hard evidence suggest that there are physical, emotional, and mental health benets conferred on people who share stories about illnesses. Further, stories have the potential to create and strengthen relationships, and also to cultivate empathy and understanding in ourselves and others who bear witness to the words and visages of suering people. Many caregivers and people with dementia have begun to communicate their stories in various forms. However, until now, we have never attempted to systematically collect and analyse stories about personal experiences of illnesses such as so-called Alzheimers disease and analyse them with the same transdisciplinary approaches to tissues that have precipitated biological insights. Our pilot banking of stories has begun in two areas: what quality of life means for people with dementia and what the perceptions of genetic risk are for those who have family members with dementia. The rst set of stories was provided by residents with a diagnosis of dementia at an assisted living facility in Cleveland. These individuals
www.thelancet.com Vol 373 April 4, 2009

volunteered to read with children at the Intergenerational Schoolan innovative charter school in Cleveland. Contrary to the expectations some might have of Alzheimers victims, the participants in the study shared a need to maintain a sense of purpose and usefulness in their lives, and expressed that the relationships they formed with children of the school met this need and contributed to their quality of life. Some even identied health benets that they attributed to working with children: a reduction in psychosocial stress and an elevation of mood. Our second set of narratives was donated by individuals at risk of dementia by virtue of a family history. They had been tested for the presence of a susceptibility gene and the quantitative research showed that they reported acceptable safety and satisfaction with the information received. Yet their narratives told a more complex story. People often did not understand, remember, or act on what was shared with them about their relative risk. This, in turn, underscores the ethical implications of disclosing vague genetic information to patients in the clinic or via directto-consumer companies. Our clinical experience and StoryBanks pilot data show that good medical care for those with cognitive decline must transcend the current reductionist approach to include and embrace the life stories of individuals. Labels need to be rethought. Alzheimers disease should be viewed, at rst, as a two-word phrase that communicates a frightening cultural story rather than a specic pathology. It is time to reconsider the scientic and cultural story of this disease and imagine new personalised narratives based on a life-long, ecological view of brain health. Our own stories about brain ageing can be informed and enriched by the stories of those with cognitive challenges and their carers. Almost 25 years ago, we began a journey with Bill shaped by a commitment to science and to care that was embodied by the gift of his mothers brain tissue. Others have enriched our understanding of the value of words and narratives in service of patients and communities. Today is the time to reassert the value of stories and appreciate the power of myths in medicine. With the innovative use of information technology, new media, and hermeneutic approaches, we can improve on the abilities that human beings have shown over the centuries to tell meaningful stories in the service of health, solidarity, and survival. It is our ambition that the concept of StoryBank can create a big humanities and social sciences initiative on a par with the big biology project of mapping the human genome. At no time in the history of our species has the need for shared storytelling and the potential resultant collective wisdom and action been greater.

Further Reading Ballenger J. Self, Senility, and Alzheimers Disease in Modern America: A History. Baltimore: Johns Hopkins University Press, 2006. Kleinman A. The Illness Narratives: Suering, Healing and the Human Condition. New York: Basic Books, 1988. Stein J, Schettler T, Rohrer B, Valenti M. Environmental Threats to Healthy Aging: With a Closer Look at Alzheimers & Parkinsons Diseases. Boston: Greater Boston Physicians for Social Responsibility and Science and Environmental Health Network, 2008. http://www. agehealthy.org/ (accessed Feb 23, 2009). Whitehouse PJ, Maurer K, Ballenger JF, eds. Concepts of Alzheimer Disease: Biological, Clinical, and Cultural Perspectives. Baltimore: Johns Hopkins University Press, 1997. Whitehouse PJ, George DR. The Myth of Alzheimers: What You Arent Being Told About Todays Most Dreaded Diagnosis. New York: St. Martins Press, 2008. http://www.themythof alzheimers.com/(accessed Feb 23, 2009). For more on the Intergenerational School see http://www.tisonline.org We would like to thank Bill Keane and the too numerous to mention other participants and faculty who contributed their stories and their expertise to our eorts, as well as the Greenwall Foundation and the National Institutes of Health, which supported some of our work.

*Peter J Whitehouse, Daniel George

peter.whitehouse@case.edu
Case Western Reserve University, Fairhill Center, Cleveland, OH 44120, USA (PJW); and Hertford College, University of Oxford, UK (DG)

1167

Obituary

Willem Johan Kol

Pioneering inventor of articial organs. Born on Feb 14, 1911, in Leiden, the Netherlands, he died on Feb 11, 2009, in Newton Square, PA, USA, aged 97 years.
Whenever I see a problem, Willem Kol once said, I try to reduce it to simple terms. If the problem is very complicated I look at whether or not there is a simple component to it. And if the simple component is an important part, I take that rst. For Kol himself this formula clearly paid o. In a long and productive working life he set up the rst blood bank in Europe, devised a cluster of medical inventions including an intra-aortic balloon pump for patients in acute cardiac failure, and, most notably, the rst articial kidneys and implantable hearts. The inventing began in the Netherlands where Kol was born to a father who ran a tuberculosis clinic. Having graduated in 1938 from the medical school at University of Leiden, Kol took a job in Groningen University. Motivated in part by having witnessed the death of a patient through renal failure he was seized by the idea of turning dialysis already a concept familiar to doctorsinto a practicable method of replacing the functions of a diseased kidney. Using various bits of available equipment, including barrels, laundry tubs, and, above all, cellophane tubing of the kind used to encase sausages, he cobbled together an articial kidney. The rst 15 of his patients survived only a few days, but the 16th lived. Kidney dialysis worked. In the Netherlands, by this time emerging from the years of war and German occupation, times were hard. In
1168

the small town hospital in Kampento which Kol had moved to avoid working under a Nazi-appointed head of departmentopportunities for medical invention were limited. But in 1950 he took a job in the USA at the Cleveland Clinic in Ohio. Here, with dialysis now established and undergoing rapid development and improvement, he set himself a new goal: to invent a heart-lung machine. The pump-oxygenator he built was eective and helped to make open heart surgery a feasible technique. But Kol had become increasing absorbed in another and even greater challenge: to create an implantable articial heart. After moving to the University of Utah, in 1967, Kol began making important progress on this new ambition. As director of its Institute of Biomedical Engineering he assembled a team of surgeons, physicists, and cardiologists. Among them was Joe Andrade who went on to become the Universitys professor of bioengineering. Kol always had the capacity to empower others, says Andrade: People from around the planet were attracted to work with him. He sometimes saw more in people than they could see in themselves. He could invigorate them to do interesting things. Many of these people went on to found their own groups and technologies. Another ex-colleague, a veterinarian by training, is professor Don Olsen who joined Kol at Utah in 1972. Although he describes his sometime employer as a demanding, and stern boss and teacher, he adds that Kol was able to see further down the road than most of us. He would keep challenging people to do something better. Kol and his team developed a series of articial hearts, each named after the member of the group responsible for that particular device. The model that eventually proved itself in patients was perfected by one of the young doctors in the team, Robert Jarvik. In an operation done in December, 1982, the rst recipient of the new heart was a retired dentist called Barney Clark. He lived for almost 4 months. Although everyone within the eld of articial organs knows the full extent of Kols contributions, some think he didnt always receive the public recognition due to him. He was generous at sharing the credit with his co-workers and that became most pronounced at Utah with the rst articial heart implantation, says Andrade. Kol encouraged his sta to put their own names to the devices they were working on. He did this to motivate people, says Olsen. So at the time of the rst implant, the media concentrated on Jarvik. Andrade recalls how He [Jarvik] happened to be young, and somewhat charismatic, and he kind of became the posterchild for the event. Andrade conrms the oft repeated suggestion that Kol was disappointed about one thing: not to have received a Nobel Prize. That said, he won pretty well everything else including, in 2002, a Lasker Award. He is survived by four sons and a daughter.

University of Utah

Geo Watts
geo@scileg.freeserve.co.uk

www.thelancet.com Vol 373 April 4, 2009

Correspondence

Lymphadenectomy in endometrial cancer

The main nding of the MRC ASTEC trial (Jan 10, p 125)1 is that there is no evidence of benet from pelvic lymphadenectomy for patients with endometrial cancer. In this trial, surgery consisted of a total abdominal hysterectomy and bilateral salpingooophorectomy with or without pelvic lymphadenectomy by laparotomy. However, surgery was also allowed by laparoscopy provided that the procedure could be accomplished safely and as thoroughly as an open procedure. However, whether laparoscopy for endometrial cancer is associated with a better or worse prognosis than open surgery is not known. Since the primary outcome of the study was whether lymphadenectomy could improve the survival of women with early endometrial cancer when compared with surgery without lymphadenectomy, we are puzzled as to why surgeons were allowed to enter patients who underwent a nonstandard surgical procedure. Although the proportion who underwent laparoscopy was similar in both groups (6%), and hence the estimate of the dierence between the groups is unlikely to be aected, the absolute gures for survival might be changed either positively or negatively. Total laparoscopic hysterectomy is not a standard surgical procedure in endometrial cancer. Laparoscopy for endometrial cancer can only be considered equivalent to an open surgical procedure when this has been proven by a randomised controlled trial. Since we are still awaiting the results of the GOG-LAP2 study in earlystage endometrial cancerthe rst randomised trial powered for survival as a primary outcomethis is not the case yet.
We declare that we have no conict of interest.

Department of Gynaecologic Oncology (MJEM, CBMB), and Epidemiology (GHdB), University Medical Center Groningen, University of Groningen, PO 30.001, 9700 RB Groningen, Netherlands 1 The writing committee on behalf of the ASTEC study group. Ecacy of systematic pelvic lymphadenectomy in endometrial cancer (MRC ASTEC trial): a randomised study. Lancet 2009; 373: 12536.

Aoun Hakmi
aoun_hakmi@hotmail.com
Conquest Hospital, St Leonards on Sea TN37 7RD, UK 1 The writing committee on behalf of the ASTEC study group. Ecacy of systematic pelvic lymphadenectomy in endometrial cancer (MRC ASTEC trial): a randomised study. Lancet 2009; 373: 12536. Mohan DS, Samuel MA, Selim MA, et al. Long-term outcomes of therapeutic pelvic lymphadenectomy for stage I endometrial adenocarcinoma. Gynecol Oncol 1998; 70: 16571. Lutman CV, Havrilesky LJ, Cragun JM, et al. Pelvic lymph node count as an important prognostic variable for FIGO stage I and II endometrial carcinoma with high-risk histology. Gynecol Oncol 2006; 102: 9297.

*M J E Mourits, C B M Bijen, G H de Bock

m.j.e.mourits@og.umcg.nl

I have some concerns about the ASTEC study.1 First, in the standard surgery group, surgeons could remove pelvic lymph nodes if they believed this to be in the womans best interest. This concession contradicts the aim of the study to test the therapeutic eect of lymphadenectomy. The nal conclusion should thus be that lymphadenectomy cannot be recommended unless the surgeon believes it is in the womens best interest (whatever is meant by that). Second, inclusion of low-risk women (43%) in the study will have diluted any possible benecial eect of lymphadenectomy in the highrisk group. Third, half the patients had 12 nodes or fewer removed (median 12). Other studies of therapeutic lymphadenectomy have a much higher median (31).2 Further studies3 have set 12 nodes as the minimum for better staging. Fourth, despite there being more women with high-risk and advanced cancer in the lymphadenectomy group, radiotherapy was still given to an equal number of women in each group, resulting in a higher proportion of the high-risk-earlystage (4%) and advanced cancer (8%) groups undergoing radiotherapy in the standard group. These were even higher5% and 9%, respectivelyfor external-beam radiation therapy. The nal results were not adjusted for this. Finally, since randomisation did not necessarily prove reliable (there was a 10% dierence in stage IC disease between groups), the second randomisation (radiotherapy) might have invalidated the results of the surgical treatments.
I declare that I have no conict of interest.

The ASTEC study group1 conclude that a systematic lymphadenectomy in endometrial cancer cannot be recommended as a routine procedure because of lack of benet in terms of recurrence-free and overall survival. However, there are several reasons why the ASTEC trial did not show improved overall survival with routine lymphadenectomy. First, the number of lymph nodes resected was insucient in many patients. Although the median number resected overall was 12, 35% of patients in the lymphadenectomy group had nine or fewer lymph nodes removed. Cragun and colleagues2 showed that removal of more than 11 pelvic nodes had an eect on overall survival. Chan and colleagues3 showed that, in intermediate-risk and high-risk endometrial cancer, patients with more than 10 nodes harvested have an improved outcome. Second, many patients with lowrisk endometrial carcinoma, and hence a low risk of lymph-node involvement, were included (eg, only 41% had stage ICIIB disease, and only 22% presented with poorly dierentiated tumours). The high rate of inclusion of low-risk patients and the low number of lymph nodes removed are the reasons for the low rate of involved lymph nodes seen in the lymphadenectomy group (9%). Third, the study group did not assess the para-aortic nodes. However, up to 67% of patients

Submissions should be made via our electronic submission system at http://ees.elsevier.com/ thelancet/

www.thelancet.com Vol 373 April 4, 2009

1169

Science Photo Library

Correspondence

with lymph-node metastases have involved para-aortic nodes.4 Fourth, the ASTEC trial was too small to detect an overall survival dierence because the expected proportion of isolated pelvic lymph-node recurrences is as low as 23% in early endometrial carcinoma. In conclusion, we believe that there is still an indication to do a comprehensive lymphadenectomy to select patients at high risk of pelvic side wall recurrence. The selection of patients for a lymphadenectomy should be based on myometrial invasion, grade, and diameter of the tumour.5
We declare that we have no conict of interest.

*Frdric Amant, Patrick Neven, Ignace Vergote

Frederic.Amant@uz.kuleuven.ac.be
Division of Gynaecological Oncology, Department of Obstetrics & Gynaecology, University Hospitals Leuven, Herestraat 49, 3000 Leuven, Belgium 1 The writing committee on behalf of the ASTEC study group. Ecacy of systematic pelvic lymphadenectomy in endometrial cancer (MRC ASTEC trial): a randomised study. Lancet 2009; 373: 12536. Cragun JM, Havrilesky LJ, Calingaert B, et al. Retrospective analysis of selective lymphadenectomy in apparent early-stage endometrial cancer. J Clin Oncol 2005; 23: 366875. Chan JK, Cheung MK, Huh WK, et al. Therapeutic role of lymph node resection in endometrioid corpus cancer: a study of 12,333 patients. Cancer 2006; 107: 182330. Mariani A, Dowdy SC, Cliby WA, et al. Prospective assessment of lymphatic dissemination in endometrial cancer: a paradigm shift in surgical staging. Gynecol Oncol 2008; 109: 1118. Amant F, Moerman P, Neven P, et al. Endometrial cancer. Lancet 2005; 366: 491505.
1

We wonder whether the stratication into risk subgroups (only briey mentioned in the Results section) has sucient statistical power for detecting a therapeutic value in highrisk patients. Second, since 67% of endometrial cancer patients with positive nodes have para-aortic involvement, a surgical procedure aimed at removing metastatic nodal disease with therapeutic intent must include bilateral systematic dissection of paraaortic nodes.2 However, para-aortic lymphadenectomy was not included in this trial. As a consequence, paraaortic residual disease was left in more than half of patients with lymphatic malignant dissemination. Third, one of the theoretical benets of surgical staging is the use of nodal information to modulate postoperative treatment,3 but the design of this trial does not allow testing of this hypothesis. Given the above observations, the conclusions of the ASTEC study should be that pelvic node dissection has no therapeutic eect in most patients with endometrial carcinoma. The real question is which subgroup of high-risk patients might possibly benet from systematic surgical staging to guide postoperative treatment.3 The question remains open.
We declare that we have no conict of interest.

Authors reply
We should emphasise that there are now two independent randomised controlled trials of lymphadenectomy in early endometrial cancer (with a total of 1922 women), both of which have conrmed no clinical benet for lymphadenectomy.1,2 In fact, in both randomised trials there is a suggestion of worse survival in the lymphadenectomy group: the hazard ratio for overall survival in ASTEC (116, 95% CI 087154)1 and Panici and colleagues (120, 070207),2 and the pooled hazard ratio combining data from both trials (117, 091150), favour the standard surgery group. M J E Mourits and colleagues ask about the use of laparoscopic surgery, which was evolving over the course of the trial. It was allowed under strict conditions and included as a stratication factor to avoid an imbalance (and bias) between the two groups. Bias from substandard surgery in the lymphadenectomy group is extremely unlikely since surgical procedures (hysterectomy and bilateral salpingo-oophorectomy) and lymph-node counts in the lymphadenectomy group were similar, if not better, in women who had laparoscopic procedures. Aoun Hakmi asks whether allowing node sampling in the standard group aected the results. The intention was to mirror clinical practice and allow surgeons to remove suspicious nodes that they were unhappy to leave in situ; 5% of patients in the standard surgery group had any nodes removed compared with 92% in the lymphadenectomy group. Turning to the reliability of the randomisation procedure, by chance there were slightly more higher-risk women in the lymphadenectomy group, which is why we did adjusted analyses to take account of these dierences. With respect to small imbalances in external-beam radiation therapy within subgroups, it is neither methodologically appropriate to adjust for post-randomisation treatments,
www.thelancet.com Vol 373 April 4, 2009

Stefano Uccella, Karl C Podratz, Giovanni D Aletti, *Andrea Mariani

mariani.andrea@mayo.edu
Division of Gynecologic Surgery, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA 1 The writing committee on behalf of the ASTEC study group. Ecacy of systematic pelvic lymphadenectomy in endometrial cancer (MRC ASTEC trial): a randomised study. Lancet 2009; 373: 12536. Mariani A, Dowdy SC, Cliby WA, et al. Prospective assessment of lymphatic dissemination in endometrial cancer: a paradigm shift in surgical staging. Gynecol Oncol 2008; 109: 1118. Mariani A, Dowdy SC, Keeney GL, Long HJ, Lesnick TG, Podratz KC. High-risk endometrial cancer subgroups: candidates for target-based adjuvant therapy. Gynecol Oncol 2004; 95: 12026.

The ASTEC study group asserts that there is no evidence of a benet for systematic lymphadenectomy in patients with endometrial cancer. This statement is subject to several pitfalls in trial design and execution. The negative results are therefore expected and not surprising. First, inclusion of patients with stage IAIB, grade 12 disease (447% of all cases) led to a low rate (9%) of nodal metastases and to surgical overtreatment of 91% of patients.
1170

Correspondence

nor necessary because it is now clear that external-beam radiation therapy has no eect on overall survival.3 Aoun Hakmi, Frdric Amant and colleagues, and Stefano Uccella and colleagues suggest that we should not have included women whose risk of nodal metastases is low. It is possible that for these women the eect of lymphadenectomy might be smaller, but an analysis of treatment eect by risk group did not provide evidence of this. The power to detect a treatment eect in the high-risk early-stage subgroup (80 deaths in 507 patients) is obviously lower than the overall power, but is similar to that in the Panici study2 (53 deaths in 514 women), which also showed no evidence of benet. Hakmi and Amant and colleagues ask about the number of nodes removed. Observational studies linking node counts to outcome are subject to bias from factors unrelated to therapeutic potential and cannot lead to reliable conclusions of treatment eect. The median number of nodes removed was used to group centres to look for an eect associated with removing more nodes: none was found. An unbiased analysis of whether removing more nodes might lead to benet on an individual patient basis is not possible. Uccella and colleagues and Amant and colleagues argue that residual unresected para-aortic disease is a reason that no benet was seen. We do not regard metastatic disease involving para-aortic nodes as being surgically curable, and with no evidence of benet from two randomised trials, it does not seem logical to suggest that an even more extensive operation (with more postoperative morbidity) should be done. Lymphadenectomy was introduced as part of International Federation of Gynecology and Obstetrics staging without evidence of benet, and it is now incumbent on those who argue for its continuation to provide this evidence. Surgical oncology merits as robust an evidence base as is expected of non-surgical cancer treatment.
www.thelancet.com Vol 373 April 4, 2009

We declare that we have no conict of interest.

Henry Kitchener, *Ann Marie Swart, Wendi Qian, Mahesh Parmar

ams@ctu.mrc.ac.uk
University of Manchester, Manchester, UK (HK); and MRC Clinical Trials Unit, London NW1 2DA, UK (AMS, WQ, MP) 1 ASTEC Writing Committee on behalf of ASTEC study group. Ecacy of systematic pelvic lymphadenectomy in endometrial cancer (MRC ASTEC trial): a randomised study. Lancet 2009; 373: 12536. Panici PB, Basile S, Maneschi F, et al. Systematic pelvic lymphadenectomy vs no lymphadenectomy in early-stage endometrial carcinoma: randomised clinical trial. J Natl Cancer Inst 2008; 100: 170716. ASTEC/EN.5 writing committee on behalf of the ASTEC/EN.5 study group. Adjuvant external beam radiotherapy in the treatment of endometrial cancer (MRC ASTEC and NCIC CTG EN.5 randomised trials): pooled results, systematic review and meta-analysis. Lancet 2009; 373: 13746.

Cytochrome P450 2C19 polymorphism and clopidogrel after MI

Jean-Philippe Collet and colleagues (Jan 24, p 309)1 show that young patients with a previous myocardial infarction who are carriers of the CYP2C19*2 allele have a signicantly worse long-term outcome than those without this allele. Previous studies have shown that patients carrying this polymorphism have reduced platelet inhibition when exposed to clopidogrel.2 Collet and colleagues correctly point out that strategies to overcome this lack of platelet inhibition, such as using higher doses of clopidogrel, are worthy of investigation. The patient group without the CYP2C19*2 allele still had an event rate of 289 per 100 patient-years for death, non-fatal myocardial infarction, or urgent revascularisation during follow-up. It would be interesting to know whether platelet function in patients who had further cardiovascular events in this group was inhibited suciently by clopidogrel. Although these patients might not be genetically predisposed to a lack

of platelet inhibition, other factors such as a lack of patients adherence to clopidogrel therapy could aect it. Spertus and colleagues3 showed that more than 13% of patients had discontinued use of clopidogrel within 30 days of having a myocardial infarction. Jackevicius and colleagues4 showed that a change in the healthcare system in Canada from prior authorisation to limited-use prescription of clopidogrel improved both uptake of clopidogrel therapy and clinical outcomes in patients after myocardial infarction. Although the trend in medicine is towards targeting therapy tailored to an individuals genetic prole, the importance of the need for improvement in the delivery of health care and of patients adherence to therapy should not be underestimated.
We declare that we have no conict of interest.

Sanjeev Bhattacharyya, *Roby Rakhit

Roby.Rakhit@royalfree.nhs.uk
Department of Cardiology, Royal Free Hospital, Pond Street, London NW3 2QG, UK
Corbis

Collet JP, Hulot JS, Pena A, et al. Cytochrome P450 2C19 polymorphism in young patients treated with clopidogrel after myocardial infarction: a cohort study. Lancet 2009; 373: 30917. Trenk D, Hochholzer W, Fromm MF, et al. Cytochrome P450 2C19681>A polymorphism and high on-clopidogrel platelet reactivity associated with adverse 1-year clinical outcome of elective percutaneous coronary intervention with drug-eluting or bare metal stents. J Am Coll Cardiol 2008; 51: 192534. Spertus JA, Kettelkamp R, Vance C, et al. Prevalence, predictors and outcomes of premature discontinuation of thienopyridine therapy after drug eluting stent placement; results from the PREMIER registry. Circulation 2006; 113: 280309. Jackevicius CA, Tu JV, Demers V, et al. Cardiovascular outcomes after a change in prescription policy for clopidogrel. N Engl J Med 2008; 359: 180210.

Jean-Philippe Collet and co-workers1 disclose that the genetic variant CYP2C19*2 is a major determinant of prognosis in young patients on clopidogrel treatment after myocardial infarction. We think that the results should be interpreted with caution.
1171

Correspondence

Clopidogrel, as a prodrug, is known to require metabolism before it can inhibit adenosine-diphosphate-induced platelet aggregation.2 Growing evidence reveals that the response to clopidogrel can be aected by pharmacokinetic variables such as intestinal absorption, metabolic activation, and biological activity, all of which are aected by genetic polymorphisms. Previously identied polymorphisms that modulate clopidogrel absorption (ABCB1), metabolic activation (CYP3A4, CYP3A5, and CYP2C19), and biological activity (P2RY12, ITGA2, and ITGB3)25 have been shown to have an eect on responsiveness to clopidogrel associated with ischaemic events. Collet and colleagues only genotyped the CYP2C19*2 variant; other functional genetic variants were not assessed. Therefore, the possibility that these genetic factors could have had an eect on the response to clopidogrel and on clinical events cannot be excluded. Whether CYP2C19*2 carriage is associated with other genetic determinants of risk also needs to be explored in further studies. Additionally, stroke is usually regarded as a clinical outcome event in most studies associated with the response to clopidogrel.4,5 However, stroke was not included in this study, which could result in an underestimate of the number of patients with adverse events. Careful consideration of these factors in this study would have made the results more credible.
We declare that we have no conict of interest.

Savi P, Pereillo JM, Uzabiaga MF, et al. Identication and biological activity of the active metabolite of clopidogrel. Thromb Haemost 2000; 84: 89196. Suh JW, Koo BK, Zhang SY, et al. Increased risk of atherothrombotic events associated with cytochrome P450 3A5 polymorphism in patients taking clopidogrel. CMAJ 2006; 174: 171522. Trenk D, Hochholzer W, Fromm MF, et al. Cytochrome P450 2C19 681G>A polymorphism and high on-clopidogrel platelet reactivity associated with adverse 1year clinical outcome of elective percutaneous coronary intervention with drug-eluting or bare-metal stents. J Am Coll Cardiol 2008; 51: 192534.

Jianting Miao, Rui Liu, *Zhuyi Li

lizhuyi@fmmu.edu.cn
Department of Neurology, Tangdu Hospital, Fourth Military Medical University, Xian City, Shaanxi Province 710038, China 1 Collet JP, Hulot JS, Pena A, et al. Cytochrome P450 2C19 polymorphism in young patients treated with clopidogrel after myocardial infarction: a cohort study. Lancet 2009; 373: 30917. Taubert D, von Beckerath N, Grimberg G, et al. Impact of P-glycoprotein on clopidogrel absorption. Clin Pharmacol Ther 2006; 80: 486501.

The study by Jean-Philippe Collet and colleagues1 does not mention the initial dose of clopidogrel, although a maintenance dose of 75 mg was considered in the inclusion criteria. This omission raises a couple of important questions. First, could increasing the dose of clopidogrel overcome the loss of eect brought about by the genetic variant? A 600 mg dose of clopidogrel used in the ARMYDA-2 trial2 showed a huge reduction in periprocedural myocardial infarction. The ongoing CURRENT/OASIS7 trial (NCT00335452)3 assessing the ecacy and safety of standard-dose and higher-dose clopidogrel will give us more denite answers. Second, if an alternative platelet inhibitor such as prasugrel, ticagrelor, or cangrelor that does not require hepatic transformation is used, would individuals with a CYP2C19*2 genetic variant have better clinical outcomes? We need a robust prospective study to treat patients, particularly those with poor metabolism of clopidogrel, and not subject patients to highcost genetic testing when we have evidence suggesting that a higher dose of clopidogrel might enhance the ecacy in a cost-eective and time-saving manner.
I declare that I have no conict of interest.

Collet JP, Hulot JS, Pena A, et al. Cytochrome P450 2C19 polymorphism in young patients treated with clopidogrel after myocardial infarction: a cohort study. Lancet 2009; 373: 30917. Patti G, Colonna G, Pasceri V, et al. Randomized trial of high loading dose of clopidogrel for reduction of periprocedural myocardial infarction in patients undergoing coronary intervention: results from the ARMYDA-2 (Antiplatelet therapy for Reduction of MYocardial Damage during Angioplasty) study. Circulation 2005; 111: 2099106. Mehta SR, Bassand JP, Chrolavicius S, et al. Design and rationale of CURRENT-OASIS 7: a randomized, 2 x 2 factorial trial evaluating optimal dosing strategies for clopidogrel and aspirin in patients with ST and non-STelevation acute coronary syndromes managed with an early invasive strategy. Am Heart J 2008; 156: 108088.

Authors reply
Sanjeev Bhattacharyya and Roby Rakhit lucidly point to the questions of pharmacodynamic response to clopidogrel (platelet function) and of compliance to treatment. Residual platelet aggregation (in response to 20 mol/L adenosine diphosphate) was actually measured by light transmission aggregometry in a subset of our population (76 of 259: 33 carriers and 43 non-carriers of the CYP2C19*2 genetic variant) and we are glad to provide these additional ndings here. Residual platelet aggregation was higher in carriers than noncarriers (411% vs 298%, p=0033) conrming the ndings of previous studies.1 Major adverse cardiovascular events occurred more frequently in carriers than in non-carriers (48% vs 7%, p<00001). Patients with major adverse cardiovascular events (n=19) tended also to have higher residual platelet aggregation than patients free of events (n=57) (416% vs 322%, p=017). There was no dierence in residual platelet aggregation between non-carriers with recurrent events and non-carriers free of event. Compliance is always dicult to assess, but in our study, patients were checked every 6 months and were censored from the analysis if clopidogrel was discontinued. Finally, because the genetic variant exerts
www.thelancet.com Vol 373 April 4, 2009

Radhakrishnan Ramaraj
drkutty2@gmail.com
Department of Internal Medicine, University of Arizona College of Medicine, 1501, N Campbell Avenue, Tucson, AZ 85724, USA

1172

Correspondence

its eect on clinical outcome in clopidogrel-treated patients, a lack of compliance with clopidogrel could only blunt the dierence between the two groups (with or without the variant) and would suggest that the eect we measured is underestimated. Jianting Miao and colleagues question other factors that could have a role in non-response to clopidogrel. The analysis was adjusted for the other known clinical or treatment factors. Concerning the other potential genetic variants that might aect clopidogrel responsiveness, the CYP2C19*2 variant has been associated with pharmacokinetic and pharmacodynamic changes in many independent populations, whereas the eect of other genetic variants has not been consistent. To avoid multiple comparisons, we used a gene candidate approach by focusing on the eect of a single genetic variant. The inuence of other genetic variants has to be tested in future studies that could use a genome-wide approach to decipher the multigenic nature of clopidogrel responsiveness. Stroke was not considered because it is a very rare event in young patients with myocardial infarction. Over-riding clopidogrel resistance is a key practical issue raised by Radhakrishnan Ramaraj, but it goes beyond the scope of our paper. Bonello and colleagues2 have shown that reloading patients with a poor pharmacodynamic response to clopidogrel can reduce the rate of poor responders. Future studies such as the ARCTIC study (NCT00827411) will provide information on the potential of individualised antiplatelet therapy based on the genetic or plateletfunctional status of the patients, but the studies cited by Ramaraj cannot answer that question.
We declare that we have no conict of interest other than that stated in the original paper.

Institut de Cardiologie, Hpital Piti-Salptrire, 47 Boulevard de lHpital, 75013 Paris, France (JPC, GM); and Service de Pharmacologie, Unit de Pharmacogntique, INSERM 621, Hpital PitiSalptrire, Paris, France (JSH) 1 Trenk D, Hochholzer W, Fromm MF, et al. Cytochrome P450 2C19 681G>A polymorphism and high on-clopidogrel platelet reactivity associated with adverse 1-year clinical outcome of elective percutaneous coronary intervention with drug-eluting or bare-metal stents. J Am Coll Cardiol 2008; 51: 192534. Bonello L, Camoin-Jau L, Arques S, et al. Adjusted clopidogrel loading doses according to vasodilator-stimulated phosphoprotein phosphorylation index decrease rate of major adverse cardiovascular events in patients with clopidogrel resistance: a multicenter randomized, prospective study. J Am Coll Cardiol 2008; 51: 140411.

Subdural haemorrhage and child maltreatment

We agree with Carole Jenny (Jan 17, p 195)1 that child abuse and neglect exists and is harmful to children and society. However, we are concerned with her recommended way to support paediatricians as they go about helping society protect children from maltreatment. Jenny insists that the nature of evidence available when diagnosing child abuse is often dierent from the evidence paediatricians use to solve other clinical dilemmas. She justies the use of observational data in the decision-making process, and condemns the basing of testimony on pet theories such as the totally discredited suggestion that hypoxia causes subdural haemorrhages. However, Jenny seems unaware of the scientic evidence. A series of 55 post-mortem examinations in fetuses and neonates2 found that hypoxia is frequently associated with a lm subdural haemorrhage arising from bleeding in the falx and tentorium cerebelli. An association between intradural haemorrhage, subdural haemorrhage, and hypoxia in newborns and infants has also been described by use of magnetic resonance with T2-weighted images at autopsy.3 A further study used

carbon dioxide asphyxiation in rats to show the occurrence of various cranial haemorrhages that were prominent in the dura,4 and another showed the presence of subdural haemorrhage and intracranial haemorrhages in children with congenital heart disease and in the setting of non-traumatic brain hypoxia.5 The nding of a subdural haemorrhage without a history or other objective evidence of trauma should not prompt the diagnosis of child abuse on its own. The fact that hypoxia has emerged as a cause of subdural haemorrhage in the context of natural conditions mandates that investigations in suspected child abuse need to be more stringent and based on evidence rather than on collections of clinical observational data.
We declare that we have no conict of interest.
Photolibrary

*Marta Cohen, Irene Scheimberg

Marta.Cohen@sch.nhs.uk
Sheeld Childrens Hospital NHS Foundation Trust, Sheeld S10 2TH, UK (MC); and Barts and the London NHS Trust, London, UK (IS) 1 Jenny C. Supporting pediatricians who work in child maltreatment. Lancet 2009; 373: 19597. Cohen MC, Scheimberg I. Evidence of occurrence of intradural and subdural hemorrhage in the perinatal and neonatal period in the context of hypoxic ischemic encephalopathy: an observational study from two referral institutions in the United Kingdom. Pediatr Dev Pathol 2008; published online Nov 13. DOI:10.2350/08-08-0509.1. Cohen MC, Whitby E. Hypoxia, intradural hemorrhage and subdural bleeding in the pediatric and perinatal post-mortem: are they related? A study combining the use of autopsy and post mortem magnetic resonance. Presented at the British Association in Forensic Medicine winter meeting; Bath, UK; Nov 24, 2007. Hauser R, Jankowski Z, Gos T, Krzyanowski M. Hemorrhages in head tissues during the asphyxiation process. Forens Sci Int 2001; 124: 23536. McQuillen PS, Barkovich AJ, Hamrick EGS, et al. Temporal and anatomic risk prole of brain injury with neonatal repair of congenital heart defects. Stroke 2007; 38: 73641.

Authors reply
Marta Cohen and Irene Scheimberg seem to have misinterpreted my remarks. They imply that I do not think scientic evidence is necessary in the courtroom. In fact, I stated that in child abuse we cannot
1173

J P Collet, J S Hulot, *G Montalescot

gilles.montalescot@psl.ap-hop-paris.fr

www.thelancet.com Vol 373 April 4, 2009

Correspondence

The printed journal includes an image merely for illustration

Panos Pictures

use controlled clinical trials in which children are prospectively randomised to those who are injured and those who are not to discern their dierences. In many areas of medicine, including Cohen and colleagues studies, observational evidence and case series based on clinical experience are relied on to provide useful data. More specically, Cohen and Scheimberg1 found subdural haemorrhages in the posterior fossa of fetuses and newborns and attributed their cause to hypoxia. How, then, does one account for the frequently found posterior fossa haemorrhages in normal, non-asphyxiated newborns?2 It might well be that subdural bleeding is part of the normal birth process. The subjects of the McQuillen3 study quoted by the correspondents were also newborns. The same common nding of subdural haemorrhage from birth could be a factor in these cases as well. More signicantly, the infants in this study had severe congenital heart disease and multiple cerebral strokes. They experienced cardiopulmonary bypass, deep hypothermia, and circulatory arresthardly a model on which to assess clinical ndings in normal infants. Certainly, cautious debate on scientic issues should respectfully continue about these important topics.
I testify in legal cases involving head trauma.

McQuillen PS, Barkovich AJ, Hamrick EGS, et al. Temporal and antomic risk prole of brain injury with neonatal repair of congenital heart defects. Stroke 2007; 38: 73641.

Make the money work for health in sub-Saharan Africa

As sub-Saharan Africa struggles to meet the Millennium Development Goals (MDGs),1 we nd your Editorial on Europes aid to the health sector in the region (Feb 7, p 434)2 very timely. Achieving the MDGs in sub-Saharan Africa will depend partly on regional policymakers accessing the best evidence about what approaches work, and integrating this evidence into national health systems.3 Thus serious consideration should be given to ensuring that funding for the health sector in sub-Saharan Africa is conditional on engagement with systematic reviews. The advantages of systematic reviews include the ability to reduce bias and chance in the assessment of existing research,4 and to provide a way for policymakers to access the global evidence on key questions in a timely way. Funding conditionality could take many forms. At a minimum, there should be clear documentation of how relevant reviews are identied and assessed for their quality, local applicability, equity impacts, and scaling-up considerations. If no reviews relevant to sub-Saharan Africa are found, funding for the health sector should include resources for high-quality monitoring and assessment of approaches and programmes. If relevant reviews are found, there should be clear documentation about whether and how the reviews informed the denition of the problem being addressed, the approach or programmes selected, and the implementation strategies planned.

Without such a systematic approach, large amounts of aid funds and local resources will continue to be squandered on ineective approaches and programmes, and millions of people will continue to endure untold and avoidable suering.
We declare that we have no conict of interest.

*Charles Shey Wiysonge, John N Lavis, Jimmy Volmink

charles.wiysonge@mrc.ac.za
South African Cochrane Centre, Medical Research Council, Cape Town, South Africa (CSW, JV); Institute of Infectious Disease and Molecular Medicine, University of Cape Town, South Africa (CSW); Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, ON, Canada (JNL); and Faculty of Health Sciences, University of Stellenbosch, Cape Town, South Africa (JV) 1 United Nations. The Millennium Development Goals Report, 2008. http:// www.un.org/millenniumgoals/pdf/The%20M illennium%20Development%20Goals%20Rep ort%202008.pdf (accessed Jan 30, 2009). The Lancet. Europes aid to sub-Saharan Africa: better results needed. Lancet 2009; 373: 434. Lavis JN, Posada FB, Haines A, Osei E. Use of research to inform public policymaking. Lancet 2004; 364: 161521. Chalmers I, Haynes B. Reporting, updating, and correcting systematic reviews of the eects of health care. BMJ 1994; 309: 86265.

Department of Error
Widmark A, Klepp O, Solberg A, et al, for the Scandinavian Prostate Cancer Group Study 7 and the Swedish Association for Urological Oncology 3. Endocrine treatment, with or without radiotherapy, in locally advanced prostate cancer (SPCG-7/ SFUO-3): an open randomised phase III trial. Lancet 2009; 373: 30108In this Article (Jan 24), the y-axis in gure 2A should read: Cumulative incidence of PSA recurrence (%). Tfelt-Hansen P. Migraine and olcegepant. Lancet 2009; 373: 1003In this Correspondence letter (March 21) the title should have read: Migraine and telcagepant. Waterston T, Halileh S, Odeh J, Rudolf M, Hamilton P. Teaching child health in the occupied Palestinian territory. Lancet 2009; 373: 87880In this Comment (March 14), the acknowledgments should have included: We thank Juzoor, the Foundation for Health and Social Development, a Palestinian nongovernmental organisation that acts as course organiser and manages the quality control for the teaching programme in the West Bank, for their support.

Carole Jenny
cjenny@lifespan.org
Warren Alpert Medical School at Brown University, Providence, RI 02903, USA 1 Cohen MC, Scheimberg I. Evidence of occurrence of intradural and subdural hemorrhage in the perinatal and neonatal period in the context of hypoxic ischemic encephalopathy: an observational study from two referral institutions in the United Kingdom. Pediatr Dev Pathol 2008; published online Nov 13. DOI:10.2350/0808-0509.1. Looney CB, Smith JK, Merck LH, et al. Intracranial hemorrhage in asymptomatic neonates: prevalence on MR images and relationship to obstetric and neonatal risk factors. Radiology 2007; 242: 53541.

1174

www.thelancet.com Vol 373 April 4, 2009

Articles

Reduction in C-reactive protein and LDL cholesterol and cardiovascular event rates after initiation of rosuvastatin: a prospective study of the JUPITER trial
Paul M Ridker, Eleanor Danielson, Francisco A H Fonseca, Jacques Genest, Antonio M Gotto Jr, John J P Kastelein, Wolfgang Koenig, Peter Libby, Alberto J Lorenzatti, Jean G MacFadyen, Brge G Nordestgaard, James Shepherd, James T Willerson, Robert J Glynn, on behalf of the JUPITER Trial Study Group

Summary
Background Statins lower high-sensitivity C-reactive protein (hsCRP) and cholesterol concentrations, and hypothesis generating analyses suggest that clinical outcomes improve in patients given statins who achieve hsCRP concentrations less than 2 mg/L in addition to LDL cholesterol less than 18 mmol/L (<70 mg/dL). However, the benet of lowering both LDL cholesterol and hsCRP after the start of statin therapy is controversial. We prospectively tested this hypothesis. Methods In an analysis of 15 548 initially healthy men and women participating in the JUPITER trial (87% of full cohort), we prospectively assessed the eects of rosuvastatin 20 mg versus placebo on rates of non-fatal myocardial infarction, non-fatal stroke, admission for unstable angina, arterial revascularisation, or cardiovascular death (prespecied endpoints) during a maximum follow-up of 5 years (median 19 years), according to on-treatment concentrations of LDL cholesterol (18 mmol/L or <18 mmol/L) and hsCRP (2 mg/L or <2 mg/L). We included all events occurring after randomisation. This trial is registered with ClinicalTrials.gov, number NCT00239681. Findings Compared with placebo, participants allocated to rosuvastatin who achieved LDL cholesterol less than 18 mmol/L had a 55% reduction in vascular events (event rate 111 vs 051 per 100 person-years; hazard ratio [HR] 045, 95% CI 034060, p<00001), and those achieving hsCRP less than 2 mg/L a 62% reduction (event rate 042 per 100 personyears; HR 038, 95% CI 026056, p<00001). Although LDL cholesterol and hsCRP reductions were only weakly correlated in individual patients (r values <015), we recorded a 65% reduction in vascular events in participants allocated to rosuvastatin who achieved both LDL cholesterol less than 18 mmol/L and hsCRP less than 2 mg/L (event rate 038 per 100 person-years; adjusted HR 035, 95% CI 023054), versus a 33% reduction in those who achieved one or neither target (event rate 074 per 100 person-years; HR 067, 95% CI 052087) (p across treatment groups <00001). In participants who achieved LDL cholesterol less than 18 mmol/L and hsCRP less than 1 mg/L, we noted a 79% reduction (event rate 024 per 100 person-years; HR 021, 95% CI 009052). Achieved hsCRP concentrations were predictive of event rates irrespective of the lipid endpoint used, including the apolipoprotein B to apolipoprotein AI ratio. Interpretation For people choosing to start pharmacological prophylaxis, reduction in both LDL cholesterol and hsCRP are indicators of successful treatment with rosuvastatin. Funding AstraZeneca.
Lancet 2009; 373: 117582 Published Online March 29, 2009 DOI:10.1016/S01406736(09)60447-5 See Comment page 1147 Center for Cardiovascular Disease Prevention (Prof P M Ridker MD, E Danielson MIA, J G MacFadyen BA, R J Glynn ScD) and Division of Cardiovascular Medicine (Prof P M Ridker, Prof P Libby MD), Brigham and Womens Hospital, Harvard Medical School, Boston, MA, USA; Universidade Federal de Sao Paulo, Sao Paulo, Brazil (Prof F A H Fonseca MD); McGill University Health Center, Montreal, Quebec, Canada (Prof J Genest MD); Weill Cornell Medical College of Cornell University, New York, NY, USA (Prof A M Gotto Jr MD); Department of Vascular Medicine, Academic Medical Center, University of Amsterdam, Amsterdam, Netherlands (Prof J J P Kastelein MD); University of Ulm Medical Center, Ulm, Germany (Prof W Koenig MD); Hospital Cordoba, Cordoba, Argentina (Prof A J Lorenzatti MD); Herlev Hospital, Copenhagen University Hospital, Herlev, Denmark (Prof B G Nordestgaard MD); University of Glasgow, Glasgow, UK (Prof J Shepherd MD); and St Lukes Episcopal Hospital/ Texas Heart Institute, Houston, TX, USA (Prof J T Willerson MD) Correspondence to: Dr Paul Ridker, Center for Cardiovascular Disease Prevention, Brigham and Womens Hospital, Boston, MA 02215, USA pridker@partners.org

Introduction
Present guidelines for statin therapy emphasise the need to achieve specic goals for LDL cholesterol to maximise clinical outcomes.1,2 However, statin therapy has greatest ecacy in the presence of inammation,3,4 and several studies show that statins reduce the inammatory biomarker high-sensitivity C-reactive protein (hsCRP) largely independently of LDL cholesterol.58 Furthermore, in both the Pravastatin or Atorvastatin Evaluation and Infection Therapy (PROVE ITTIMI 22)9 and Aggrastat-to-Zocor (A to Z)10 trials of patients with acute coronary ischaemia treated with statin therapy, the best clinical outcomes were in those who not only achieved LDL cholesterol less than 18 mmol/L (<70 mg/dL), but who also achieved hsCRP less than 2 mg/L. These ndings are consistent with the pathophysiological
www.thelancet.com Vol 373 April 4, 2009

understanding that atherothrombosis is a disorder of both hyperlipidaemia and inammation11 and that statins have anti-inammatory and lipid-lowering properties.12,13 Despite the consistency of these results, whether achieving lower concentrations of hsCRP after initiation of statin therapy is associated with improved clinical outcomes, in a similar manner to that associated with achieving lower concentrations of LDL cholesterol, remains controversial. We prospectively tested this hypothesis in the large-scale JUPITER (Justication for the Use of statins in Prevention: an Intervention Trial Evaluating Rosuvastatin) trial.14

Methods
Patients and procedures
The study population was derived from JUPITERa randomised, double-blind, placebo-controlled trial that

1175

Articles

Placebo (N=7832)

Rosuvastatin LDL 18 mmol/L (N=2110) LDL <18 mmol/L (N=5606) 660 (610710) 2108 (376%) 285 (255321) 135 (124146) 80 (7588) 814 (145%) 657 (117%) 2337 (420%) 42 (2868) 27 (2431) 13 (1015) 134 (096191) 34 (3038) 107 (094120) 07 (0608) 522 (488566) 57% (5459) hsCRP 2 mg/L (N=4305) 660 (610710) 1798 (418%) 290 (257330) 135 (125146) 80 (7588) 740 (172%) 473 (110%) 1859 (435%) 54 (3686) 28 (2431) 13 (1015) 136 (099192) 35 (3138) 109 (095122) 07 (0508) 522 (483566) 57% (5560) hsCRP <2 mg/L (N=3411) 660 (600710) 1149 (337%) 277 (249309) 134 (123145) 80 (7587) 452 (133%) 421 (124%) 1279 (378%) 32 (2447) 28 (2531) 13 (1116) 130 (093188) 34 (3138) 109 (095122) 07 (0508) 522 (488566) 57% (5459) hsCRP <2 mg/L+LDL hsCRP <1 mg/L+LDL <18 mmol/L (N=2685) <18 mmol/L (N=944) 660 (600710) 916 (341%) 278 (252311) 134 (124146) 80 (7587) 329 (123%) 342 (128%) 1036 (388%) 32 (2446) 28 (2431) 13 (1015) 132 (094189) 34 (3038) 107 (094121) 07 (0508) 522 (488566) 57% (5459) 650 (600710) 276 (292%) 274 (246305) 132 (120144) 80 (7486) 115 (122%) 116 (124%) 331 (363%) 28 (2242) 28 (2430) 13 (1016) 129 (092184) 34 (3037) 108 (094120) 07 (0508) 522 (488561) 56% (5459)

Age (years) Women Body-mass index (kg/m2) Blood pressure (mm Hg) Systolic Diastolic Current smoking Family history of coronary disease Metabolic syndrome* hsCRP (mg/L) LDL cholesterol (mmol/L) HDL cholesterol (mmol/L) Triglycerides (mmol/L) Non-HDL cholesterol (mmol/L) Apolipoprotein B (g/L) Apolipoprotein B to apolipoprotein AI ratio Glucose (mmol/L) HbA1c (%)

660 (600710) 2957 (378%) 284 (253320) 134 (124145) 80 (7587) 1231 (157%) 936 (120%) 3274 (421%) 43 (2871) 28 (2431) 13 (1016) 133 (097191) 35 (3138) 109 (096122) 07 (0608) 522 (488-566) 57% (5559)

650 (600710) 839 (398%) 278 (248315) 134 (124144) 80 (7587) 378 (179%) 237 (113%) 801 (383%) 42 (2871) 29 (2631) 13 (1116) 130 (095189) 35 (3239) 113 (100127) 07 (0608) 516 (477555) 57% (5559)

Data are median (IQR) or number (%). hsCRP=high-sensitivity C-reactive protein. HbA1c=haemoglobin A1c. *Metabolic syndrome was dened according to consensus criteria of the American Heart Association and the National Heart, Lung, and Blood Institute.15

Table 1: Baseline clinical characteristics of the study population in the placebo and rosuvastatin groups according to achieved concentrations of LDL cholesterol and high-sensitivity C-reactive protein (hsCRP)

was designed to investigate whether rosuvastatin 20 mg daily decreased the rate of rst cardiovascular events compared with placebo in apparently healthy men and women with LDL cholesterol less than 337 mmol/L (<130 mg/dL) who are at increased vascular risk due to
Events/ patients Target concentration Placebo Rosuvastatin LDLC 18 mmol/L LDLC <18 mmol/L p value Percentage reduction Placebo Rosuvastatin LDLC reduction <50% 65/4181 LDLC reduction 50% 38/3535 p value 074 047 070 (053093) 041 (029059) <00001 070 (053093) 042 (030059) <00001 066 (050088) 042 (029059) <00001 189/7832 111 100 100 100 39/2110 64/5606 091 051 089 (063125) 045 (034060) <00001 089 (063126) 046 (035061) <00001 085 (060121) 045 (033059) <00001 189/7832 111 100 100 100 Event HRAge (95% CI) rate* HRAge,LDL,CRP (95% CI) HRFully adjusted (95% CI)

HR=hazard ratio. Data include all events occurring after randomisation. *Event rates are per 100 person-years. Fully adjusted model controlled for age, baseline LDL cholesterol, baseline high-sensitivity C-reactive protein (hsCRP), baseline HDL cholesterol, blood pressure, sex, body-mass index, smoking status, and parental history of premature coronary heart disease.

Table 2: Hazard ratios for incident cardiovascular events in the JUPITER trial according to magnitude of reduction in LDL cholesterol (LDLC) in participants allocated to rosuvastatin

hsCRP concentration of 2 mg/L or more. Full details of the trial protocol and procedures have been previously presented.14 The prespecied primary endpoint of non-fatal myocardial infarction, non-fatal stroke, admission for unstable angina, arterial revascularisation, or cardiovascular death was used for all analyses contained in this study. To assess the eect of reductions in hsCRP and LDL cholesterol on trial event rates, on an a priori basis, and as part of the study protocol, concentrations of hsCRP and LDL cholesterol were obtained at randomisation and every year thereafter. For this analysis, we used baseline and 1-year values for hsCRP and LDL concentrations, which were available for 15 548 trial participants (87% of the full JUPITER cohort). Since statins maximally reduce LDL cholesterol and hsCRP within 6 weeks, we decided a priori to include all events arising after randomisation in our primary analyses instead of arbitrarily limiting the analysis to events that occurred after any specic timepoint. This approach is conservative since any misclassication of on-treatment concentrations of LDL cholesterol or hsCRP on this basis would, if anything, bias the analysis towards a null result. To test the validity of these assumptions, we also undertook secondary analyses limited only to events that occurred after the rst year of treatment.
www.thelancet.com Vol 373 April 4, 2009

1176

Articles

Statistical analysis
We used Spearman correlation coecients to assess the relation between concentrations of on-treatment LDL cholesterol and on-treatment hsCRP, as well as the correlation between percentage change over 1 year in these two variables. Consistent with the PROVE ITTIMI 22 and A to Z trials,9,10 we rst divided the participants given rosuvastatin into two groups on the basis of whether or not the achieved LDL cholesterol concentration was 18 mmol/L or more, or less than this value, and used Cox proportional-hazards models to estimate relative hazards for cardiovascular events in these two groups, compared with participants allocated placebo. We also divided the participants given rosuvastatin into two groups on the basis of whether or not the achieved hsCRP concentration was greater than 2 mg/L, or less than this value, and used Cox proportional-hazards models to estimate relative hazards for cardiovascular events in these two groups, compared with those allocated placebo. As an internal check to ensure the validity of this a-priori approach, we repeated the above process dividing the cohort on the basis of achieving reductions in LDL cholesterol of greater than or less than 50%, and on the basis of achieving reductions in hsCRP of greater than or less than 50%. To address the eect of the achieved hsCRP concentrations across the strata of achieved LDL cholesterol values, we further repeated this process after dividing the cohort given rosuvastatin into four groups on the basis of LDL cholesterol cuto of 18 mmol/L, and hsCRP cuto of 2 mg/L. We did a test for trend in event rates across these groups by assigning a score of 0 to the placebo group, a score of 1 to those in the rosuvastatin group who achieved neither target (LDL cholesterol 18 mmol/L, CRP 2 mg/L), a score of 2 to those in the rosuvastatin group who achieved one target (LDL cholesterol 18 mmol/L, CRP <2 mg/L; or LDL cholesterol <18 mmol/L, CRP 2 mg/L), and a score of 3 to those in the rosuvastatin group who achieved both targets (LDL cholesterol <18 mmol/L, CRP <2 mg/L). For ease of clinical application, we undertook similar analyses after dividing the participants given rosuvastatin into two groups: those who achieved LDL cholesterol less than 18 mmol/L and CRP less than 2 mg/L, and those who did not achieve one or both of these targets. We also compared outcomes for participants achieving or not achieving each of the above LDL cholesterol and hsCRP targets, restricting the analysis to those allocated to active rosuvastatin treatment. Since previous retrospective analyses9,10 raised the hypothesis of greatest benets in people who not only achieved LDL cholesterol less that 18 mmol/L, but who also achieved hsCRP concentrations less than 1 mg/L, we also undertook analyses with these alternative predetermined targets. To address whether alternative lipid measures might aect results, we repeated all analyses substituting non-HDL cholesterol, apolipoprotein B, or the ratio of
www.thelancet.com Vol 373 April 4, 2009
N Placebo LDL cholesterol achieved 18 mmol/L LDL cholesterol achieved <18 mmol/L Placebo LDL cholesterol reduction <50% LDL cholesterol reduction 50% Placebo hsCRP achieved 2 mg/L hsCRP achieved <2 mg/L Placebo hsCRP reduction <50% hsCRP reduction 50% 7832 2110 5606 7832 4181 3535 7832 4305 3411 7832 4143 3573 Rate 111 091 051 111 074 047 111 077 042 111 070 051 025 05 10

p<00001

p<00001

p<00001

p<00001

20

40

Rosuvastatin better

Rosuvastatin worse

Figure 1: Hazard ratios for incident cardiovascular events in JUPITER according to achieved concentrations of LDL cholesterol or high-sensitivity C-reactive protein (hsCRP) after initiation of rosuvastatin Data are adjusted for age, baseline LDL and HDL cholesterol, baseline hsCRP, blood pressure, sex, body-mass index, smoking status, and parental history of premature coronary heart disease. Event rates are per 100 person-years. Events/ patients Target concentration Placebo Rosuvastatin hsCRP 2 mg/L hsCRP <2 mg/L p value Percentage reduction Placebo Rosuvastatin hsCRP reduction <50% hsCRP reduction 50% p value 63/4143 40/3573 070 051 064 (048085) 046 (033065) <00001 064 (048085) 046 (033065) <00001 063 (047084) 044 (031062) <00001 189/7832 111 100 100 100 72/4305 31/3411 077 042 069 (053091) 038 (026056) <00001 069 (053090) 038 (026056) <00001 068 (051089) 036 (024054) <00001 189/7832 111 100 100 100 Event rate* HRAge (95% CI) HRAge,LDL,CRP (95% CI) HRFully adjusted (95% CI)

HR=hazard ratio. Data include all events occurring after randomisation. *Event rates are per 100 person-years. Fully adjusted model controlled for age, baseline LDL cholesterol, baseline hsCRP, baseline HDL cholesterol, blood pressure, sex, body-mass index, smoking status, and parental history of premature coronary heart disease.

Table 3: Hazard ratios for incident cardiovascular events in the JUPITER trial according to magnitude of reduction in high-sensitivity C-reactive protein (hsCRP) in participants allocated to rosuvastatin

apolipoprotein B to apolipoprotein AI for LDL cholesterol. Further sensitivity analyses were done by varying target lipid concentrations. This trial is registered with ClinicalTrials.gov, number NCT00239681.

Role of the funding source

JUPITER was an investigator-initiated trial. The sponsor of the study collected the trial data and monitored the study sites, but had no role in the conduct of the analyses or drafting of the report. All statistical analyses were done by the investigators and the academic study statistician
1177

Articles

Events/ patients hsCRP cuto <2 mg/L Placebo Rosuvastatin LDL cholesterol 18 mmol/L, hsCRP 2 mg/L LDL cholesterol >18 mmol/L, hsCRP < 2 mg/L LDL cholesterol <18 mmol/L, hsCRP 2 mg/L LDL cholesterol <18 mmol/L, hsCRP <2 mg/L p value Rosuvastatin LDL cholesterol 18 mmol/L or hsCRP 2 mg/L LDL cholesterol <18 mmol/L and hsCRP <2 mg/L p value hsCRP cuto <1 mg/L Placebo Rosuvastatin LDL cholesterol 18 mmol/L, hsCRP 1 mg/L LDL cholesterol 18 mmol/L, hsCRP <1 mg/L LDL cholesterol <18 mmol/L, hsCRP 1 mg/L LDL cholesterol <18 mmol/L, hsCRP <1 mg/L p value Rosuvastatin LDL cholesterol 18 mmol/L or hsCRP 1 mg/L LDL cholesterol <18 mmol/L and hsCRP <1 mg/L p value 98/6772 5/944 36/1874 3/236 189/7832 80/5031 31/1384 8/726 41/2921 189/7832

Event rate* 111 111 054 062

HRAge (95% CI)

HRAge,LDL,CRP (95% CI) 100 107 (073156) 056 (028114) 054 (039076) 036 (023055) <00001 067 (052087) 036 (023055) <00001 100 093 (065133) 067 (021210) 050 (038067) 022 (009054) <00001 061 (048078) 022 (009054) <00001

HRFully adjusted (95% CI) 100 106 (072155) 042 (018094) 053 (038074) 035 (023054) <00001 064 (049084) 035 (023054) <00001 100 089 (062128) 046 (011185) 049 (037066) 021 (009051) <00001 059 (046075) 021 (009051) <00001

100 106 (072155) 054 (027110) 055 (039077) 035 (023054) <00001

23/2685 038

074

067 (052087) 035 (023054) <00001

23/2685 038

111 095 064

100 091 (064130) 065 (021203) 050 (038067) 021 (009052) <00001

59/4662 056 5/944 024

067 024

061 (048077) 021 (009052) <00001

HR=hazard ratio. Data include all events occurring after randomisation.*Event rates are per 100 person-years. Fully adjusted model controlled for age, baseline LDL cholesterol, baseline hsCRP, baseline HDL cholesterol, blood pressure, sex, body-mass index, smoking status, and parental history of premature coronary heart disease.

Table 4: Hazard ratios for incident cardiovascular events in the JUPITER trial in the placebo group and according to achieved concentrations of both LDL cholesterol and high-sensitivity C-reactive protein (hsCRP) in participants allocated to rosuvastatin

(RJG). PMR and RJG had full access to all study data and had nal responsibility for the decision to submit for publication.

Results
Table 1 shows baseline characteristics of the study population in the placebo and rosuvastatin groups according to achieved LDL cholesterol and achieved hsCRP concentrations. Baseline age, blood pressure, HDL cholesterol, triglycerides, glucose, and haemoglobin A1c were much the same between groups (table 1). As expected, baseline LDL cholesterol values were lower in particpants given rosuvastatin who subsequently
1178

achieved LDL cholesterol less than 18 mmol/L compared with those who did not; similarly baseline hsCRP concentrations were lower in those given rosuvastatin who achieved hsCRP less than 2 mg/L (table 1). Participants achieving lower LDL cholesterol and hsCRP concentrations on rosuvastatin were less likely to be women or to smoke, had lower body-mass index, and were more likely to have a family history of coronary disease (table 1). All these potential confounding factors, as well as baseline HDL cholesterol and blood pressure, were included in the fully adjusted models. By contrast, in participants allocated to placebo, only 106 (1%) achieved the targets of LDL cholesterol less than 18 mmol/L and hsCRP less than 2 mg/L, and 29 (<1%) achieved the targets of LDL cholesterol less than 18 mmol/L and hsCRP less than 1 mg/L, largely because of o-trial statin use. Rosuvastatin reduced the median LDL cholesterol by 50% (p<00001) and the median hsCRP by 37% compared with placebo (p<00001). However, the magnitude of correlation between achieved LDL cholesterol and achieved hsCRP in participants given rosuvastatin was small (r=010), so that less than 2% of the variance in achieved hsCRP was explained by the variance in achieved LDL cholesterol. The correlations between achieved hsCRP and achieved apolipoprotein B (r=010), between achieved hsCRP and the achieved ratio of apolipoprotein B to apolipoprotein AI (r=016), and between the percentage reduction in hsCRP and the percentage reduction in LDL cholesterol (r=015) were also small. As previously reported,14 rosuvastatin allocation resulted in a 44% reduction in the trial primary endpoint (HR 056, 95% CI 046069, p<00001). We detected no evidence of a signicant interaction between the overall ecacy of rosuvastatin and baseline concentrations of hsCRP less than 5 mg/L (HR 049, 95% CI 037065) or 5 mg/L or greater (066, 049089; p for interaction=015), or baseline LDL cholesterol less than 26 mmol/L (066, 047092) or 26 mmol/L or more (052, 040067; p for interaction=028). However, as expected, the magnitude of reduction in LDL cholesterol with rosuvastatin was directly related to the magnitude of clinical benet. Compared with placebo, participants allocated to rosuvastatin who did not achieve LDL cholesterol less than 18 mmol/L had no signicant reduction in vascular events (HR 089, 95% CI 063125, p=049), whereas we recorded a 55% reduction in those who did achieve this target (045, 034060, p<00001) (p for trend across LDL cholesterol strata <00001, p for comparison between active treatment groups=0001). These eects were minimally changed after adjustment for baseline LDL cholesterol or for any of the other risk factors at study entry (table 2). We recorded a similar relation between on-treatment LDL cholesterol and event rates in analyses stratied by percentage reduction (table 2, gure 1).
www.thelancet.com Vol 373 April 4, 2009

Articles

N Placebo LDL cholesterol 18 mmol/L, hsCRP 2 mg/L LDL cholesterol <18 mmol/L, hsCRP 2 mg/L LDL cholesterol 18 mmol/L, hsCRP <2 mg/L LDL cholesterol <18 mmol/L, hsCRP <2 mg/L Placebo LDL cholesterol 18 mmol/L, hsCRP 1 mg/L LDL cholesterol <18 mmol/L, hsCRP 1 mg/L LDL cholesterol 18 mmol/L, hsCRP <1 mg/L LDL cholesterol <18 mmol/L, hsCRP <1 mg/L 7832 1384 2921 726 2685 7832 1874 4662 236 944

Rate 111 111 062 054 038 111 095 056 064 024 025 05 10 20

p<00001

p<00001

40

Rosuvastatin better

Rosuvastatin worse

Figure 2: Hazard ratios for incident cardiovascular events in JUPITER according to achieved concentrations of LDL cholesterol and high-sensitivity C-reactive protein (hsCRP) after initiation of rosuvastatin Data are adjusted for age, baseline LDL and HDL cholesterol, baseline hsCRP, blood pressure, sex, body-mass index, smoking status, and parental history of premature coronary heart disease. Event rates are per 100 person-years.

Despite little correlation between reductions in LDL cholesterol and hsCRP, the magnitude of decrease in hsCRP was directly related to the magnitude of clinical benet. Compared with placebo, participants allocated to rosuvastatin who did not achieve hsCRP less than 2 mg/L had a 31% decrease in events (HR 069, 95% CI 053091, p=0007), whereas we noted a 62% reduction in those who did achieve this hsCRP target (038, 026056, p<00001) (p for trend across hsCRP strata <00001, p for comparison between active treatment groups=0007). These eects were minimally changed after adjustment for baseline hsCRP or for any of the other risk factors at baseline (table 3). We recorded a similar relation between on-treatment hsCRP and event rates in analyses stratied by percentage reduction (table 3, gure 1). We detected the lowest risk of cardiovascular events in participants given rosuvastatin who not only achieved low concentrations of LDL cholesterol, but who also achieved low values of hsCRP (table 4 and gure 2). In similar analyses restricted to participants who achieved LDL cholesterol concentrations less than 18 mmol/L, participants who achieved hsCRP concentrations less than 2 mg/L had better clinical outcomes than did those who did not (table 4). Figure 3A shows the cumulative incidence of cardiovascular events in the placebo and rosuvastatin groups, according to whether or not the targets of LDL cholesterol less than 18 mmol/L and hsCRP less than 2 mg/L were achieved. In fully adjusted analyses, the hazard ratio was 064 (95% CI 049084) for one or both targets not achieved, and 0.35 (023054) for both targets achieved (p for trend across groups <00001, p for comparison between active treatment groups 0033). We recorded a 79% reduction in hazard in participants who achieved the targets of LDL cholesterol less than 18 mmol/L and hsCRP less than 1 mg/L (table 4,
www.thelancet.com Vol 373 April 4, 2009

gure 2). In similar analyses restricted to participants who achieved LDL cholesterol concentrations less than 18 mmol/L, those who achieved hsCRP concentrations less than 1 mg/L had better clinical outcomes than did those who did not. Figure 3B shows the cumulative incidence of cardiovascular events in the placebo and rosuvastatin groups, according to whether or not the targets of LDL cholesterol less than 18 mmol/L and hsCRP less than 1 mg/L were achieved. In fully adjusted analyses, the hazard ratio was 059 (95% CI 046075) for one or both targets not achieved, and 021 (009051) for both targets achieved (p for trend across groups <00001, p for comparison between active treatment groups=0037). We did several additional analyses to address the robustness of these ndings. First, although our primary a-priori analyses were based on all events after randomisation with achieved concentrations of LDL cholesterol and hsCRP available, we repeated these analyses limiting the data to events occurring after the 1-year blood samples were obtained, and noted a very similar pattern of results (webappendix p 1). Second, because others have postulated that non-HDL cholesterol, apolipoprotein B, or the ratio of apolipoprotein B to apolipoprotein AI might be better than LDL cholesterol for monitoring of statin therapy,16,17 we repeated these analyses substituting each of these alternative lipid measures for LDL cholesterol. For these analyses, we used a non-HDL cholesterol target of less than 26 mmol/L, an apolipoprotein B target of less than 08 g/L, and a target ratio of apolipoprotein B to apolipoprotein AI of less than 05, since the proportions of participants above and below these cutos were similar to that above or below the target of LDL cholesterol less than 18 mmol/L. In all these analyses, participants achieving low concentrations of hsCRP and low values of the alternative lipid variable had better

See Online for webappendix

1179

Articles

A
008 Placebo Rosuvastatin (LDL cholesterol 18 mmol/L or hsCRP 2 mg/L) Rosuvastatin (LDL cholesterol <18 mmol/L and hsCRP <2 mg/L)

006 Cumulative incidence

004

18 mmol/L. In this sensitivity analysis, participants who achieved the targets of LDL cholesterol less than 14 mmol/L and hsCRP less than 2 mg/L had better clinical outcomes than did those who did not (data not shown). We recorded similar ndings in sensitivity analyses that used the median on-treatment cutos observed within the trial for non-HDL cholesterol (<20 mmol/L), apolipoprotein B (<066 g/L), or the ratio of apolipoprotein B to apolipoprotein AI (<04) (data not shown).

002

Discussion
In healthy men and women starting rosuvastatin therapy in the JUPITER trial, achievement of target concentrations of LDL cholesterol less than 18 mmol/L and hsCRP less than 2 mg/L was associated with improved event-free survival compared with achievement of neither target or with achievement of reduced LDL cholesterol alone. The dierential outcomes that we recorded on the basis of achieved concentrations of LDL cholesterol and hsCRP remained signicant and unchanged in magnitude after adjustment for all available baseline characteristics that varied between groups, including concentrations of both LDL cholesterol and hsCRP before randomisation. We detected similar eects in analogous assessments when non-HDL cholesterol, apolipoprotein B, or the ratio of apolipoprotein B to apolipoprotein AI were substituted for LDL cholesterol, and in sensitivity analyses based on even more aggressive lipid targets. We noted the greatest hazard reduction in participants who achieved LDL cholesterol concentrations less than 18 mmol/L and hsCRP concentrations less than 1 mg/L. We believe these ndings to be of interest for several reasons. First, these prospective data in an outpatient prevention setting lend support to controversial data from the PROVE ITTIMI 229 and A to Z10 trials in which patients with acute coronary syndrome who achieved hsCRP concentrations less than 2 mg/L and LDL cholesterol concentrations less than 18 mmol/L had the best clinical outcomes for patients being treated with statin therapy. The JUPITER design allowed us to simultaneously adjust for a wide range of baseline clinical characteristics, including entry values of LDL cholesterol and hsCRP. Thus, we can conclude that the dierences in clinical outcome result from drug response rather than from dierence in baseline clinical characteristics. Second, data from this study derive from a population of healthy men and women and are thus free of potential confounding eects that might have accrued in the PROVE ITTIMI 22 and A to Z studies in which on-treatment hsCRP and LDL cholesterol concentrations were ascertained after an acute ischaemic event. The present data are also consistent with intravascular ultrasound evidence from the REVERSAL trial18 of patients with stable coronary artery disease, in which
www.thelancet.com Vol 373 April 4, 2009

0 Number at risk Rosuvastatin Placebo

7716 7832

7699 7806

7678 7777

6040 6114

3608 3656

1812 1863

1254 1263

913 905

508 507

145 168

B
008 Placebo Rosuvastatin (LDL cholesterol 18 mmol/L or hsCRP 1 mg/L) Rosuvastatin (LDL cholesterol <18 mmol/L and hsCRP <1 mg/L)

006 Cumulative incidence

004

002

0 0 Number at risk Rosuvastatin Placebo 7716 7832 7699 7806 1 7678 7777 6040 6114 2 Follow-up (years) 3608 3656 1812 1863 3 1254 1263 913 905 4 508 507 145 168

Figure 3: Cumulative incidence of cardiovascular events in JUPITER in the placebo and rosuvastatin groups according to whether or not reductions in both LDL cholesterol and high-sensitivity C-reactive protein (hsCRP) were achieved (A) Analysis using targets of LDL cholesterol less than 18 mmol/L and hsCRP less than 2 mg/L. (B) Analysis using targets of LDL cholesterol less than 18 mmol/L and hsCRP less than 1 mg/L.

clinical outcomes than did those who did not achieve the respective targets (all p values for trend across groups <0001, all p values for comparisons between active treatment groups <005 apart from apolipoprotein B, for which p=0057) (gure 4). Further, irrespective of the alternative lipid measure used, we detected a similar pattern of risk reduction associated with achieved hsCRP across strata of achieved lipid concentrations (webappendix p 2). Finally, to address whether the use of alternative lipid cutos might aect our ndings, we repeated the above analyses with an LDL cholesterol target of less than 14 mmol/L (<55 mg/dL) (the median on-treatment LDL cholesterol concentrations within JUPITER) rather than the clinical target of LDL cholesterol less than
1180

Articles

plaque regression after statin therapy was detected only when both hsCRP and LDL cholesterol were reduced. Third, these data could provide insight into mechanisms by which statin therapy reduces vascular risk. In addition to being highly eective agents to reduce cholesterol through inhibition of the hydroxymethyl-glutarylcoenzyme A (HMG-CoA) reductase pathway, statins reduce inammatory cell adhesion and monocyte recruitment to endothelial cells, change smooth muscle migration in developing plaques, and favourably aect matrix metalloproteinases leading to plaque stabilisation.12,13 The extent to which these anti-inammatory properties aect clinical outcomes, and whether or not these eects are independent of LDL cholesterol, remains an intense area of research.19 Nonetheless, our previous work with pravastatin,5,6,9 lovastatin,4 cerivastatin,7 simvastatin,8,10 atorvastatin,9 and now rosuvastatin has consistently shown that more ecacious statins have a greater eect on inammation, and that for any given statin, hsCRP response varies with dose. We chose to use rosuvastatin 20 mg when designing JUPITER since a core scientic goal was to assess the use of robust LDL cholesterol and hsCRP reduction in a population of patients who did not qualify for statin therapy according to conventional guidelines, but who were at increased vascular risk because of an enhanced inammatory prole. Thus, the nding that clinical benets are maximised when both LDL cholesterol and hsCRP are reduced provides new insight into why more potent statins (as tested in the PROVE ITTIMI 22 trial)20 or higher doses of the same statin (as tested in the Treating to New Targets trial)21 result in increased clinical benets. Limitations of our data merit consideration. Although our data derive from a large randomised trial, analyses according to achieved LDL cholesterol and achieved hsCRP are no longer randomised; thus, although we adjusted for all available clinical characteristics between study groups, residual confounding cannot be fully excluded. However, our targets for achieved LDL cholesterol and achieved hsCRP were prespecied on the basis of previous published work and are not derived from data. Long-term safety data for rosuvastatin are scarce in our trial because the median exposure is 19 years (maximum 5 years). Finally, although restriction of the range of both LDL cholesterol and hsCRP by design in our study could in theory limit generalisation, these same features greatly reduce the potential for baseline confounding and thus enhance validity. Moreover, our data are consistent with those from the AFCAPS/TexCAPS4 trial that had a much wider range of concentrations of LDL cholesterol and hsCRP at baseline, and with those from the aforementioned PROVE ITTIMI 22,9 A to Z,10 and REVERSAL18 trials that studied patients with overt hyperlipidaemia. Despite the pathophysiological evidence presented here, in low-risk primary prevention populations with
www.thelancet.com Vol 373 April 4, 2009

A
008 Placebo Rosuvastatin (apolipoprotein B 08 g/L or hsCRP 2 mg/L) Rosuvastatin (apolipoprotein B <08 g/L and hsCRP <2 mg/L)

006 Cumulative incidence

004

002

0 Number at risk Rosuvastatin Placebo

7665 7832

7649 7806

7629 7777

5994 6114

3581 3656

1796 1863

1244 1263

906 905

504 507

144 168

B
008 Placebo Rosuvastatin (apolipoprotein B:apolipoprotein AI 05 or hsCRP 2 mg/L) Rosuvastatin (apolipoprotein B:apolipoprotein AI <05 and hsCRP <2 mg/L)

006 Cumulative incidence

004

002

0 0 Number at risk Rosuvastatin Placebo 7665 7832 7649 7806 1 7629 7777 5994 6114 2 Follow-up (years) 3581 3656 1796 1863 3 1244 1263 906 905 4 504 507 144 168

Figure 4: Cumulative incidence of cardiovascular events in JUPITER in the placebo and rosuvastatin groups according to whether or not reductions in alternative lipid concentrations and high-sensitivity C-reactive protein (hsCRP) were achieved (A) Analysis using targets of apolipoprotein B less than 08 g/L and hsCRP less than 2 mg/L. (B) Analysis using targets of ratio of apolipoprotein B to apolipoprotein AI less than 05 and hsCRP less than 2 mg/L.

raised LDL cholesterol or hsCRP, initial interventions should remain lifestyle recommendations for dietary restriction, exercise, and smoking cessation. However, as our ndings have shown, for people choosing to start pharmacological prophylaxis, reductions in both LDL cholesterol and hsCRP are indicators of the success of treatment with statin therapy.9,10.22
Contributors PMR, the Principal Investigator and Trial Chairman of JUPITER, designed and undertook the trial, interpreted the data, and wrote this report. RJG, the academic study statistician, along with ED and JGM, managed the dataset and undertook the independent statistical analyses. FAHF, JG, AMG, JJPK, WK, PL, AJL, BGN, JS, and JTW are members of the JUPITER Steering Committee and assisted in multiple phases of the trial including participant recruitment, data collection, and data interpretation. All authors have seen and approved the nal version of the report.

1181

Articles

Conict of interest statement During the period of this study, PMR reports having received investigator-initiated research grant support from the National Heart Lung and Blood Institute, the National Cancer Institute, the Donald W Reynolds Foundation, the Leducq Foundation, AstraZeneca, Novartis, Merck, Abbott, Roche, and Sano-Aventis; consulting fees and/or lecture fees from AstraZeneca, Novartis, Merck-Schering Plough, Sano-Aventis, ISIS, and Vascular Biogenics; and is listed as a co-inventor on patents held by the Brigham and Womens Hospital that relate to the use of inammatory biomarkers in cardiovascular disease. These patents have been licensed to several entities, including AstraZeneca. FAHF reports having received research grants, lecture fees, and consulting fees from AstraZeneca, Pzer, Schering-Plough, Sano-Aventis, and Merck. JG reports having received lecture fees from AstraZeneca, Schering Plough, Pzer, Novartis, and Sano-Aventis; and consulting fees from AstraZeneca, Merck, Schering-Plough, Pzer, Novartis, and Sano-Aventis. AMG reports having received consulting fees from Dupont, Novartis, Aegerion, Arisaph, KOWA, Merck, Merck Schering Plough, Pzer, and Reliant. JJPK reports receiving research grant support from AstraZeneca, Pzer, Roche, Novartis, Merck, Merck Schering Plough, ISIS, Genzyme, and Sano-Aventis; lecture fees from AstraZeneca, GlaxoSmithKline, Pzer, Novartis, and Boehringer-Ingelhiem; and consulting fees from AstraZeneca, Abbott, Pzer, ISIS, Genzyme, Roche, Novartis, Merck, Merck Schering Plough, and Sano-Aventis. WK reports receiving research grant support from Dade-Behring and GlaxoSmithKline; lecture fees from AstraZeneca, Pzer, Novartis, GlaxoSmithKline, and Boehringer-Ingelheim; and consulting fees from GlaxoSmithKline and Roche. PL reports consulting fees from and serves on the Scientic Advisory boards of VIA Pharmaceutical, Interleukin Genetics, Bind Biosciences, Carolus Therapeutics, and Kowa Research Institute. AJL reports receiving research grant support, lecture fees, and consulting fees from AstraZeneca, Takeda, and Novartis. BGN reports receiving lecture fees from AstraZeneca, Sano-Aventis, Pzer, Boehinger Ingelheim, and Merck; and consulting fees from AstraZeneca, Abbott, and BG Medicine. JS reports receiving lecture fees from AstraZeneca, Pzer, and Merck; and consulting fees from AstraZeneca, Merck, Pzer, Nicox, and Oxford Biosciences. RJG reports receiving research grant support from AstraZeneca and Bristol Myers Squibb. JTW, ED, and JGMF declare that they have no conict of interest. Acknowledgments JUPITER was supported by AstraZeneca, USA. We thank the medical teams from AstraZeneca whose eorts made this trial possible, the JUPITER study participants, and the more than 1000 physicians worldwide for their personal time, eort, and commitment to the JUPITER trial. References 1 De Backer G, Ambosioni E, Borch-Johnson K, et al. European guidelines on cardiovascular disease prevention in clinical practice. Eur Heart J 2003; 24: 160110. 2 Grundy SM, Cleeman JI, Merz CN, et al. Implications of recent clinical trials for the National Cholesterol Education Program Adult Treatment Panel III guidelines. Circulation 2004; 110: 22739. 3 Ridker PM, Rifai N, Pfeer MA, et al. Inammation, pravastatin, and the risk of coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events (CARE) Investigators. Circulation 1998; 98: 83944. 4 Ridker PM, Rifai N, Cleareld M, et al. Measurement of C-reactive protein for the targeting of statin therapy in the primary prevention of acute coronary events. N Engl J Med 2001; 344: 195965.

9 10

11 12 13

14

15

16

17

18

19

20

21

22

Albert MA, Danielson E, Rifai N, Ridker PM. Eect of statin therapy on C-reactive protein levels: the Pravastatin Inammation/CRP Evaluation (PRINCE), a randomized trial and cohort study. JAMA 2001; 286: 6470 Ridker PM, Rifai N, Pfeer MA, Sacks F, Braunwald E. Long-term eects of pravastatin on plasma concentration of C-reactive protein. The Cholesterol and Recurrent Events (CARE) Investigators. Circulation 1999; 100: 23035. Ridker PM, Rifai N, Lowenthal SP. Rapid reduction in C-reactive protein with cerivastatin among 785 patients with primary hypercholesterolemia. Circulation 2001; 103: 119193. Jialal I, Stein D, Balis D, Grundy SM, Adams-Huet B, Deveraj S. Eect of hydroxymethyl glutaryl coenzyme A reductase inhibitor therapy on high sensitive C-reactive protein levels. Circulation 2001; 103: 193335. Ridker PM, Cannon CP, Morrow D, et al. C-reactive protein levels and outcomes after statin therapy. N Engl J Med 2005; 352: 2028. Morrow DA, de Lemos JA, Sabatine MS, et al. Clinical relevance of C-reactive protein during follow-up of patients with acute coronary syndromes in the Aggrastat-to-Zocor Trial. Circulation 2006; 114: 28188. Hansson GK, Libby P. The immune response in atherosclerosis: a double-edged sword. Nat Rev Immunol 2006; 6: 50819. Davignon J. Benecial cardiovascular pleiotropic eects of statins. Circulation 2004; 109 (suppl 1): III-3943. Arnaud C, Braunersreuther V, Mach F. Toward immunomodulatory and anti-inammatory properties of statins. Trends Cardiovasc Med 2005; 15: 20206. Ridker PM, Danielson E, Fonseca FAH, et al, for the JUPITER Study Group. Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein. N Engl J Med 2008; 359: 2195207. Grundy SM, Cleeman JI, Daniels SR, et al. Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart Lung and Blood Institute scientic statement: executive summary. Circulation 2005; 112: 273552. Kastelein JJP, van der Steeg W, Holme I, et al, for the TNT and IDEAL Study Groups. Lipids, apolipoproteins, and their ratios in relation to cardiovascular events with statin treatment. Circulation 2008; 117: 300209. van der Steeg WA, Boekholdt AM, Stein EA, et al. Role of the apolipoporteinB-apolipoprotein A-1 ratio in cardiovascular risk assessment: a case-control analysis in EPIC-Norfolk. Ann Intern Med 2007; 146: 64048. Nissen SE, Tuzcu EM, Schoenhagen P, et al. Statin therapy, LDL cholesterol, C-reactive protein, and coronary artery disease. N Engl J Med 2005; 352: 2938. Arnaud C, Burger F, Steens S, et al. Statins reduce interleukin-6-induced C-reactive protein in human hepatocytes: new evidence for direct anti-inammatory eects of statins. Arterioscler Thromb Vasc Biol 2005; 25: 123136. Cannon CP, Braunwald E, McCabe CH, et al, for the PROVE-IT TIMI-22 Investigators. Comparison of intensive and moderate lipid lowering with statins after acute coronary syndromes. N Engl J Med 2004; 350: 1495504. LaRosa JC, Grundy SM, Waters DD, et al. Intensive lipid lowering with atorvastatin in patients with stable coronary artery disease. N Engl J Med 2005; 352: 142535. Sabatine MS, Morrow DA, Jablonski KA, et al, for the PEACE Investigators. Prognostic signicance of the Centers for Disease Control/American Heart Association high-sensitivity C-reactive protein cut points for cardiovascular and other outcomes in patients with stable coronary artery disease. Circulation 2007; 115: 152836.

1182

www.thelancet.com Vol 373 April 4, 2009

Articles

Moxioxacin versus ethambutol in the initial treatment of tuberculosis: a double-blind, randomised, controlled phase II trial
Marcus B Conde, Anne Efron, Carla Loredo, Gilvan R Muzy De Souza, Nadja P Graa, Michelle C Cezar, Malathi Ram, Mohammad A Chaudhary, William R Bishai, Afranio L Kritski, Richard E Chaisson

Summary
Background New treatments are needed to shorten the time required to cure tuberculosis and to treat drug-resistant strains. The uoroquinolone moxioxacin is a promising new agent that might have additive activity to existing antituberculosis agents. We assessed the activity and safety of moxioxacin in the initial stage of tuberculosis treatment. Methods We undertook a phase II, double-blind, randomised controlled trial of a regimen that included moxioxacin in adults with sputum smear-positive tuberculosis at one hospital in Rio de Janeiro, Brazil. 170 participants received isoniazid, rifampicin, and pyrazinamide at standard doses and were assigned by permuted block randomisation to receive either moxioxacin (400 mg) with an ethambutol placebo (n=85) or ethambutol (1520 mg/kg) plus moxioxacin placebo (n=85) 5 days per week for 8 weeks. The primary endpoint was the proportion of patients whose sputum culture had converted to negative by week 8. Analysis was by modied intention to treat (ITT); patients whose baseline cultures were negative, contaminated, or contained drug-resistant Mycobacterium tuberculosis were excluded from the analysis. Additionally, all missing 8-week results were deemed treatment failures. This study is registered with ClinicalTrials.gov, number NCT00082173. Findings 74 patients assigned to the moxioxacin group and 72 in the ethambutol group were included in the modied ITT population. 125 patients had 8-week data (moxioxacin n=64, ethambutol n=61); the main reason for absence of data was culture contamination. At 8 weeks, culture conversion to negative had occurred in 59 (80%) of 74 patients in the moxioxacin group compared with 45 (63%) of 72 in the ethambutol group (dierence 172%, 95% CI 28317; p=003). There were 16 adverse events (eight in each group) in 12 patients. Only one event was judged related to study drug (grade 3 cutaneous reaction in the ethambutol group). Interpretation Moxioxacin improved culture conversion in the initial phase of tuberculosis treatment. Trials to assess whether moxioxacin can be used to shorten the duration of tuberculosis treatment are justied. Funding US Food and Drug Administration Oce of Orphan Product Development, with additional support from the US National Institutes of Health.
Lancet 2009; 373: 118389 See Editorial page 1145 See Comment page 1148 Instituto de Doencas do Torax/ Hospital Clementino Fraga Filho, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil (M B Conde MD, C Loredo RN, G R Muzy De Souza MD, N P Graa MD, M C Cezar MD, A L Kritski MD); Center for Tuberculosis Research, Johns Hopkins University School of Medicine, Baltimore, MD, USA (A Efron MSN, Prof W R Bishai MD, Prof R E Chaisson MD); and Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA (M Ram PhD, M A Chaudhary PhD) Correspondence to: Prof Richard E Chaisson, Center for TB Research, 1550 Orleans Street, 1M.08 Baltimore, MD 21231, USA rchaiss@jhmi.edu

Introduction
The development of new drug regimens for tuberculosis is an urgent global health priority.1 Although so-called short-course treatment can eectively cure drugsusceptible tuberculosis in 6 months, a large proportion of patients in whom tuberculosis is diagnosed do not complete a course of treatment.2 New drugs that shorten the duration of tuberculosis treatment would substantially reduce the likelihood of disease recurrence and death caused by inadequate therapy. Additionally, since every year there are 500 000 reported cases of tuberculosis caused by strains of Mycobacterium tuberculosis that are resistant to the key rst-line drugs isoniazid and rifampicin, agents that are active in multidrug-resistant tuberculosis are also needed.3 Moxioxacin is a uoroquinolone that has potent in-vitro activity against M tuberculosis.4,5 Early studies of moxioxacin in murine models of tuberculosis showed that the drug had good bactericidal activity that was
www.thelancet.com Vol 373 April 4, 2009

additive to isoniazid.6,7 On the basis of these studies, we designed a phase II clinical trial to test the hypothesis that the substitution of ethambutol, an agent used primarily to prevent the emergence of drug resistance, with moxioxacin would signicantly increase the proportion of patients with negative sputum cultures after 8 weeks of treatment. 8-week sputum conversion has proved a useful surrogate marker of the sterilising activity of tuberculosis regimens,8 and a substantial improvement in this endpoint after moxioxacin treatment would support progression to a phase III trial to assess whether a regimen including this drug could shorten the duration of tuberculosis treatment.

Methods
Patients
We undertook a single-centre, randomised, double-blind, double-dummy trial of moxioxacin versus ethambutol in patients with sputum smear-positive tuberculosis
1183

Articles

197 patients assessed for eligibility 27 ineligible for enrolment 170 enrolled and randomised

Ethics Committee of the Ministry of Health (CONEP), Brasilia, and the Agencia Nacional de Vigilancia Sanitaria (ANVISA), Brazil.

Procedures
Patients were stratied by HIV serostatus and assigned by permuted block randomisation with blocks of four to receive treatment with either moxioxacin 400 mg with an ethambutol placebo or ethambutol 1520 mg/kg plus moxioxacin placebo (control). Treatment allocation was concealed and allocation slips sealed in opaque envelopes that were opened after enrolment. The study statistician, who generated the randomisation sequence, had no involvement in the conduct of the study. The study was designed to ensure that all patients received an eective regimen for tuberculosis irrespective of the contribution of the two experimental agents; therefore, all patients received isoniazid 300 mg, rifampicin 450 mg (weight <50 kg) or 600 mg (weight >50 kg), and pyrazinamide 2025 mg/kg by directly observed treatment. Patients, study clinicians, and study sta were unaware of the treatment allocation, with the exception of the pharmacist who dispensed medication packets. All treatment was given by direct supervision either in the hospital clinic or in the community by study personnel, 5 days per week. At the end of 8 weeks, all patients were placed on open-label treatment with isoniazid and rifampicin two times per week to complete another 4 months of treatment. Moxioxacin and matching placebo were provided by Bayer Healthcare (West Haven, CT, USA) and ethambutol, ethambutol placebo, isoniazid, rifampicin, and pyrazinamide were provided by Farmanguinhos (Rio de Janeiro, Brazil). All study agents were produced under good manufacturing processes. Patients visited the clinic once a week to assess clinical status and to monitor for adverse reactions. Aminotransferase concentration, serum creatinine concentration, and complete blood counts were measured every month, and an electrocardiogram was obtained at weeks 2, 4, 6, and 8. A sputum specimen (spontaneous or induced) was obtained every week for smear and culture. Sputum specimens were digested and decontaminated by the N-acetyl-L-cysteine-sodium hydroxide method.9,10 Pellets were resuspended in a nal volume of 2 mL and used immediately for inoculation of Lwenstein-Jensen culture medium. Lwenstein-Jensen slants were prepared from a commercial powder medium base (BD, Sparks, MD, USA), according to the manufacturers instructions. 200 L of each decontaminated respiratory specimen was inoculated onto each of two slants. Slants were incubated at 37C in ambient carbon dioxide and examined visually for growth twice a week for 8 weeks by laboratory personnel who were not involved in the study and blinded to treatment status. Growth on slants was quantied by number of visible colonies, and then examined by acid-fast staining to conrm the presence of mycobacteria; isolates were
www.thelancet.com Vol 373 April 4, 2009

85 assigned to moxioxacin

85 assigned to ethambutol

11 excluded 5 baseline resistance 6 baseline culture negative, contaminated, or contained non-tuberculous mycobacteria

13 excluded 6 baseline resistance 7 baseline culture negative, contaminated, or contained non-tuberculous mycobacteria

74 in modied ITT population 10 without 8-week culture data 1 withdrew 3 moved/lost to follow-up 3 culture contaminated 2 no sputum 1 died

72 in modied ITT population 11 without 8-week culture data 3 withdrew 4 culture contaminated 2 no sputum 1 died 1 drug-resistant M tuberculosis

Figure 1: Trial prole ITT=intention-to-treat.

with no previous history of treatment in Hospital Clementino Fraga Filho, Rio de Janeiro, Brazil. Patients aged 18 years or older were eligible for enrolment if they had clinical signs and symptoms of pulmonary tuberculosis, including an abnormal chest radiograph and at least one sputum smear with acid-fast bacilli visible by Ziehl-Neelsen staining. Potential participants underwent baseline screening that included blood cell counts, liver-function testing, kidney-function testing, HIV serology, chest radiography, and culture and drug susceptibility testing for mycobacteria. Exclusion criteria were: haemoglobin concentration less than 70 g/L; aspartate aminotransferase or alanine aminotransferase concentration more than three times the upper limit of normal; serum creatinine concentration more than twice the upper limit of normal; an electrocardiogram with a QTc interval more than 450 ms; pregnancy or breastfeeding; silicotuberculosis; a history of severe adverse reactions to uoroquinolone drugs or any other study agent; and seropositivity for HIV with a CD4-cell count less than 200 cells per L, since this nding would be an indication for antiretroviral therapy, which could be aected by interactions with antituberculosis drugs. Randomised patients were excluded if their baseline culture did not grow M tuberculosis or grew a strain of M tuberculosis that was resistant to isoniazid, rifampicin, or ethambutol. Written informed consent was obtained from study participants before enrolment. Ethical review of the protocol was obtained from the Johns Hopkins Medicine Institutional Review Board and the Ethics Committee of the Federal University of Rio de Janeiro, the National
1184

Articles

speciated by use of standard biochemical tests.10 Drug susceptibility testing of all baseline cultures growing M tuberculosis was done by the proportions method.10 The primary endpoint of the study was the proportion of patients with negative sputum cultures after 8 weeks of treatment. Patients were followed up for at least 12 months after completion of their 6-month course of treatment.

Moxioxacin (n=74) Age (years) Female White HIV seropositive Current or past smoker Duration of treatment before enrolment (days) Socioeconomic status Income (R$ per month) Formal employment Weight (kg) Baseline sputum acid-fast bacilli smear <1 per high-power eld 110 per high-power eld >10 per high-power eld Baseline radiograph Inltrate Bilateral disease Cavitation Adenopathy Pleural eusion Baseline symptoms Productive cough Haemoptysis Fever Night sweats Weight loss Baseline blood results White blood cells (per mL) Packed-cell volume (%) Haemoglobin (g/L) Neutrophils (per mL) Alkaline phosphatase (IU/L) Alanine aminotransferase (IU/L) Aspartate aminotransferase (IU/L) Total bilirubin (mol/L) Gamma-glutamyl transpeptidase (IU/L) Creatinine (mol/L)
Data are n (%) or mean (SD).

Ethambutol (n=72) 357 (120) 22 (31%) 32 (44%) 2 (3%) 37 (51%) 56 (25) 847 (724) 41 (57%) 582 (99) 33 (46%) 17 (24%) 22 (31%) 72 (100%) 31 (43%) 41 (57%) 11 (15%) 7 (10%) 57 (79%) 20 (28%) 58 (81%) 57 (79%) 70 (97%) 9728 (3921) 369 (51) 123 (18) 7072 (3398) 1212 (492) 356 (189) 268 (144) 84 (44) 961 (893) 707 (88)

325 (117) 34 (46%) 30 (41%) 3 (4%) 29 (39%) 64 (32) 636 (550) 35 (47%) 543 (96) 42 (57%) 10 (14%) 22 (30%) 74 (100%) 34 (46%) 59 (80%) 8 (11%) 4 (5%) 59 (80%) 25 (34%) 59 (80%) 61 (82%) 66 (89%) 9324 (2961) 364 (57) 123 (18) 6784 (2530) 1164 (623) 358 (223) 278 (163) 94 (60) 726 (505) 707 (177)

Statistical analysis
We assumed that the conversion rate in the control group (ethambutol) would be 65%,8 and that moxioxacin would lead to an increase of 20% to a rate of 85%. To achieve a power of 08 with an alpha of 005, we needed 70 patients with 8-week data in each group of the trial. We aimed to enrol 85 patients per group to allow for exclusions based on negative initial culture, growth of non-tuberculous mycobacteria, or baseline drug resistance. The primary endpoint of the trial, culture conversion, was assessed by modied intention-to-treat (ITT) analysis; patients whose baseline cultures were negative, contaminated, or contained drug-resistant M tuberculosis were excluded and all missing 8-week results were deemed treatment failures. We also analysed all patients with available 8-week cultures. p values were calculated by test. Multivariate logistic regression was done with treatment allocation and selected baseline characteristics as covariates and culture conversion as the dependent variable. Time to conversion of sputum cultures to negative was estimated by the Kaplan-Meier method and the dierence in median times compared by the log-rank test. All analyses were done with SAS version 9. This study is registered with ClinicalTrials.gov, number NCT00082173.

Role of the funding source

The funding agencies were not involved in the design, conduct, or analysis of the study. The Investigational New Drug Oce of the US FDA reviewed the protocol and made suggestions about the study design before initiation of the trial, but this was unrelated to any funding decisions. The decision to publish was made solely by the authors and the funding agencies did not review or approve the manuscript. Bayer Healthcare donated moxioxacin and matching placebo for the trial, but had no input into the study design, execution, or analysis. The corresponding author had full access to all data in the study.

Table 1: Baseline characteristics of study participants

Results
From October, 2004, up to March, 2007, 197 patients were screened for participation in the trial. Figure 1 shows the trial prole. 74 patients assigned to the moxioxacin group and 72 assigned to the ethambutol group were assessed for the primary outcome. In the modied ITT analysis, all missing data were deemed treatment failures.
www.thelancet.com Vol 373 April 4, 2009

Baseline characteristics of the patients are shown in table 1. A higher proportion of patients assigned to moxioxacin had cavitation on chest radiograph than those assigned to ethambutol. The mean baseline weight was about 4 kg lower in the moxioxacin group than in the ethambutol group. 59 (80%) of 74 patients in the moxioxacin group and 45 (63%) of 72 in the ethambutol group had negative sputum cultures at week 8 (dierence 172%, 95% CI 28317; p=003). Of patients with sputum culture data available at week 8, 59 (92%) of 64 had negative cultures in the moxioxacin group compared with 45 (74%) of 61 controls (dierence 184%, 56313; p=0006).
1185

Articles

100 90 Proportion of patients with negative culture (%) 80 70 60 50 40 30 20 10 0 0 Moxioxacin (n/N [%]) Ethambutol (n/N [%]) Dierence (% [95% CI])

Moxioxacin Ethambutol

1 9/69 (130%) 2/68 (29%) 101% (12 to 190)

2 17/70 (243%) 2/66 (30%) 213% (104 to 321)

4 5 Week of treatment 37/69 (536%) 20/65 (308%) 228% (66 to 391) 35/61 (574%) 21/64 (328%) 246% (76 to 415)

6 45/65 (692%) 40/66 (606%) 86% (76 to 249)

7 48/62 (774%) 42/66 (636%) 138% (18 to 294)

8 59/64 (922%) 45/61 (738%) 184% (56 to 313)

0/74 0/72 0 ..

31/67 (463%) 11/66 (167%) 296% (147 to 446)

Figure 2: Proportion of patients with negative sputum cultures during 8 weeks of treatment Data are for patients with 8-week culture data available at each timepoint.

Patients assigned to moxioxacin became culture negative more rapidly than those assigned to ethambutol. After only 1 week, nine (13%) of 69 patients in the moxioxacin group had negative sputum cultures compared with two (3%) of 68 in the ethambutol group (p=003). At every week after enrolment, patients assigned to moxioxacin had a higher rate of culture conversion than those assigned to ethambutol, and this dierence was signicant at all timepoints apart from weeks 6 and 7 (gure 2). The median time to consistently negative cultures was 350 days (IQR 220550) for patients in the experimental group compared with 485 days (410560) for controls (log rank p=0005). Univariate and multivariate logistic regression analysis with the modied ITT population or with those patients with sputum culture data available at week 8 (as treated) showed similar results; the modied ITT analysis is
Univariate analysis Patients with positive culture (n=42)* Age (years) Weight (kg) Female Cavitation on radiograph <1 110 >10 Moxioxacin group 356 (319392) 573 (542603) 14 (33%) 30 (71%) 15 (36%) 12 (29%) 15 (36%) 15 (36%)

shown in table 2. Treatment with moxioxacin was associated with an increase in the odds of being culture negative at week 8 in the multivariate analysis (p=00009). Age was associated with culture conversion in the univariate analysis; patients with positive cultures were a mean of 25 years older than those who had conversion to negative culture, but this association was not signicant in the multivariate analysis. Weight, sex, and cavitation on chest radiograph were not signicantly associated with the likelihood of conversion. Patients who had positive sputum cultures at 8 weeks were more likely than those with negative cultures to have had baseline sputum smears with more than ten bacilli per high-power eld. Adverse events did not dier by treatment group. There were 16 serious adverse events (eight in each group) in 12 patients (table 3); one grade 3 cutaneous reaction in the ethambutol group was judged to be related to study drugs by the treating physicians who were not aware of treatment assignment. All other serious adverse events were judged not related to study drugs. Eight patients died during the study, including one in each group still receiving study phase treatment. No death was attributed to study treatment. Only ve patients discontinued treatment because of toxic eects; two patients in the moxioxacin group stopped because of grade 2 nausea and vomiting and one because of grade 2 paraesthesias and ataxia. Two patients in the ethambutol group stopped because of grade 2 rash and pruritis and one because of grade 3 peripheral neuropathy. No clinically or statistically signicant changes in the QTc interval were recorded in patients in either group of the trial. Seven patients (5%) had recurrence of tuberculosis conrmed by positive culture and compatible clinical symptoms: three patients in the moxioxacin group (at 11, 16, and 27 months after completing treatment) and four in the ethambutol group (at 6, 7, 22, and 32 months after completion). Six of seven isolates were tested for drug resistance, and all remained susceptible to isoniazid
Multivariate analysis

Patients with negative culture (n=104)* 331 (308354) 555 (537574) 42 (40%) 70 (67%) 58 (56%) 14 (13%) 32 (31%) 59 (57%)

OR (95% CI) 098 (097100) 099 (098101) 114 (081158) 119 (084169) 075 (046125) 048 (033069) 186 (135255)

p value 003 036 045 033 027 <00001 00001

OR (95% CI) 099 (097100) 100 (098102) 105 (074151) 107 (078149) 090 (055146) 048 (033068) 175 (126244)

p value 008 087 077 067 067 <00001 00009

Acid-fast bacilli per high-power eld on smear

OR=odds ratio. *Data are mean (95% CI) or number (%). Moxioxacin versus ethambutol.

Table 2: Univariate and multivariate logistic regression analysis for culture conversion at week 8 in the modied intention-to-treat population

1186

www.thelancet.com Vol 373 April 4, 2009

Articles

and rifampicin. DNA ngerprinting data on these isolates to distinguish relapse from re-infection are not available.

Event Ethambutol Patient 1 Patient 2 Patient 3 Patient 4 Patient 5 Patient 5 Patient 6 Patient 6 Moxioxacin Patient 7 Patient 7 Patient 8 Patient 9 Patient 10 Patient 11 Patient 11 Patient 12 Hospital admission Hospital admission Death Hospital admission Death Hospital admission Death Grade 4 toxicity Hospital admission Death Death Death Hospital admission Death Hospital admission Death

Days from enrolment to event 68 268 65 31 171 240 84 95 31 377 242 57 372 19 61 69

Primary diagnosis

Grade

Discussion
In this phase II clinical trial, we found that compared with ethambutol, moxioxacin increased the proportion of sputum cultures that converted to negative in patients with pulmonary tuberculosis. More patients in the moxioxacin group were culture negative after 1 week of treatment than in the ethambutol group, and this dierence continued through the 8 weeks of intensive-phase treatment. Compared with the control group, patients assigned to moxioxacin had an absolute dierence in culture conversion after 8 weeks of just under 20%. The addition of rifampicin to tuberculosis regimens in the 1970s increased culture conversion rates at 2 months by 1520% and allowed the duration of treatment to be reduced from 18 months to 69 months.11,12 In subsequent years, the addition of pyrazinamide to regimens containing isoniazid, rifampicin, and ethambutol or streptomycin increased the 8-week sputum conversion rate by 13% and allowed treatment duration to be shortened from 9 months to 6 months.13 Our data suggest that the improvement in sterilisation provided by moxioxacin could shorten tuberculosis treatment by one or several months, on the basis of reductions in treatment duration achieved by similar improvements in 2-month culture conversion with pyrazinamide.13 Additionally, in studies published since this trial was initiated, moxioxacin has shortened treatment to 3 or 4 months in murine models of tuberculosis.14,15 Two other trials done at the same time as this study have shown potency of moxioxacin. The Tuberculosis Trials Consortium Study 27 compared moxioxacin with ethambutol and found that culture conversion occurred more rapidly in patients treated with moxioxacin; however, the proportions of patients with negative cultures were similar in both trial groups after 8 weeks.16 Study 27 was a multicentre trial and had a factorial design in which half of all patients were treated with a thrice-weekly regimen and the other half with daily treatment. Additionally, the culture methods used were not standardised in the participating laboratories, and more sensitive liquid culture media were used. We used the well-validated Lwenstein-Jensen culture method in our study, and all cultures were done in one laboratory with a standard protocol. Although a single-centre study is more ecient and ensures standardisation of laboratory methods in a phase II trial, the results from one population might not be generalisable to all populations. However, our study was done in a setting with high tuberculosis burden, and previous studies of tuberculosis treatment in similar settings have proved to be broadly relevant to the treatment of patients worldwide.1113
www.thelancet.com Vol 373 April 4, 2009

Cutaneous reaction Gunshot wound Gunshot wound Tuberculosis Polyneuropathy Unknown Unilateral leg oedema Subdural haemorrhage Community-acquired pneumonia Pulmonary abscess Urinary sepsis Spontaneous abortion Gunshot wound Dysphasia Oesophageal neoplasm Proteinuria

3 5 5 5 3 5 2 5 3 2 5 NA 5 4 5 4

NA=not applicable. All adverse events were judged as not related to study drugs, apart from the event in patient 1, which was judged as probably related to study drugs.

Table 3: Serious adverse events

The OFLOTUB study, undertaken at several sites in South Africa, found that patients treated with moxioxacin or gatioxacin had higher culture conversion rates on solid media after 8 weeks than patients treated with ethambutol, whereas conversion rates for ooxacin treatment did not dier from those for ethambutol.17 Modelled data from quantitative cultures also suggested that moxioxacin resulted in more rapid culture conversion, although solid media data were not reported for earlier timepoints. In multivariate analysis, assignment to moxioxacin was associated with culture conversion. Age was associated with culture conversion in the univariate analysis only. Older age has been associated with poorer treatment responses in several trials of tuberculosis treatment.16,18 More patients assigned to moxioxacin had cavities on chest radiograph at baseline, but this feature was not associated with lower ecacy. By contrast, several studies have reported lower culture conversion rates in patients with cavitary lesions on chest radiograph, although results from another large clinical trial did not support these ndings.16,1820 If the presence of cavities is truly associated with slower time to culture conversion, the results of this study might underestimate the potency of moxioxacin. Notably, baseline sputum-smear grade was signicantly associated with culture conversion, and heavy smear positivity (more than ten organisms per high-power eld) was equally distributed in the two study groups. Both cavitation on radiograph and sputum-smear positivity are semiquantitative measures that are subject
1187

Articles

to observer bias and inter-observer variability. The number of patients with HIV infection, which was found to be associated with poorer outcomes in the Tuberculosis Trials Consortium study,16 was too small for meaningful analysis in this study. Moxioxacin was well tolerated and not associated with increased occurrence of adverse reactions or clinical complications compared with ethambutol, a drug generally considered to be very safe. In particular, we found no change in the QTc interval on serial electrocardiograms taken during the trial (data not shown). QTc prolongation is a reported toxic eect of other uoroquinolone drugs, but has not been noted with moxioxacin during short durations of treatment. Because tuberculosis treatment lasts longer than treatment of community-acquired pneumonia (the main clinical use for moxioxacin), the nding of few adverse events is reassuring and suggests that moxioxacin could be safely used for longer periods; however, this suggestion needs to be conrmed in additional studies that are large enough to detect small but important rates of toxic eects. The results of our trial have substantial implications for future trials. First, the improved culture conversion rates found after 8 weeks in the experimental group suggest that moxioxacin, in combination with other rst-line antituberculosis drugs, could shorten the time needed to cure tuberculosis by several months. Because treatment default is directly related to the duration of treatment, a reduction in the duration of tuberculosis therapy would substantially improve outcomes. Additionally, shorter regimens for tuberculosis treatment would reduce workloads for overburdened tuberculosis control programmes, especially in high-incidence countries. Second, the demonstration of moxioxacins antimycobacterial activity shows that this agent can be used to treat tuberculosis caused by organisms with resistance to rst-line antituberculosis agents, such as isoniazid and rifampicin. In view of its clinical eectiveness and superior in-vitro potency against M tuberculosis compared with other uoroquinolone drugs,4 moxioxacin might become the preferred uoroquinolone for treatment of multidrug-resistant tuberculosis; gatioxacin might be as active, but is associated with a risk of dysglycaemia when used to treat community-acquired pneumonia in elderly patients.21 This phase II trial was intended to test the hypothesis that moxioxacin would increase the rate of sputumculture conversion after 8 weeks, and does not prove the ecacy of use of moxioxacin to shorten tuberculosis treatment. Nevertheless, our data add to a growing body of evidence that suggests that moxioxacin could shorten tuberculosis treatment by initially eradicating a greater number of organisms and improving the sterilising activity of combination drug regimens.1416,2225 Trials in animal models also suggest that replacement

of isoniazid with moxioxacin shortens tuberculosis treatment more than does replacement of ethambutol,14 but human studies have not yet supported this nding.26 The results of our study support the undertaking of clinical trials to assess whether shorter courses of moxioxacin-containing regimens can cure tuberculosis as well as or better than the current 6-month regimen. Such trials are under way and their results are eagerly awaited.
Contributors WRB, ALK, and REC designed the study. MBC, AE, CL, GRMDS, NPG, MCC, ALK, and REC participated in the implementation of the trial. MBC, AE, CL, GRMDS, and ALK participated in administration and MBC and AE in regulatory support. AE, CL, MR, MAC, and REC participated in data management. CL, NPG, and MCC participated in data collection. AE, MR, MAC, and REC contributed to data analysis. REC wrote the report. All authors saw and approved the nal report. Conict of interest statement We declare that we have no conict of interest. Acknowledgments The study was funded by the Oce of Orphan Product Development, US Food and Drug Administration (grant number R01FD002135). Additional support for training was provided by the Fogarty International Center of the US National Institutes of Health (grant TW006885 and NIH grant 01607). Moxioxacin and matching placebo were supplied by Bayer Healthcare, which had no role in the design, conduct, or analysis of the study. A career development grant helped to support RECs participation. We thank the members of our data safety and monitoring board, Reynaldo Dietze (Chair), Leda Jamal, and Ronir Raggio Luiz. We also thank Jacques Grosset, Eric Nuermberger, Andrew Vernon, Fernanda Mello, Susan Dorman, and Richard OBrien for encouragement and advice; Anna Grazia Marsico and Gisele Betzler de Oliveira Vieira for managing laboratory specimens; and Bonnie S King for assistance with data management. Finally, we thank all the patients who volunteered to participate in this trial. References 1 Young DB, Perkins MD, Duncan K, Barry CE. Confronting the scientic obstacles to global control of tuberculosis. J Clin Invest 2008; 118: 125565. 2 Dye C, Watt CJ, Bleed D. Low access to a highly eective therapy: a challenge for international tuberculosis control. Bull World Health Organ 2002; 80: 3744. 3 WHO. Global tuberculosis control 2008surveillance, planning, nancing. Geneva: World Health Organization, 2008. WHO/HTM/ TB/2008.393. 4 Ji B, Lounis N, Maslo C, Truot-Pernot C, Bonnafous P, Grosset J. In vitro and in vivo activities of moxioxacin and clinaoxacin against Mycobacterium tuberculosis. Antimicrob Agents Chemother 1998; 42: 206669. 5 Gillespie SH, Billington O. Activity of moxioxacin against mycobacteria. J Antimicrob Chemother 1999; 44: 39395. 6 Miyazaki E, Miyazaki M, Chen JM, Chaisson RE, Bishai WR. Moxioxacin (BAY12-8039), a new 8-methoxyquinolone, is active in a mouse model of tuberculosis. Antimicrob Agents Chemother 1999; 43: 8589. 7 Lounis N, Bentoucha A, Truot-Pernot C, et al. Eectiveness of once-weekly rifapentine and moxioxacin regimens against Mycobacterium tuberculosis in mice. Antimicrob Agents Chemother 2001; 45: 348286. 8 Mitchison DA. Assessment of new sterilizing drugs for treating pulmonary tuberculosis by culture at 2 months. Am Rev Respir Dis 1993; 147: 106263. 9 Kubica GP, Dye WE, Cohn ML, Middlebrook G. Sputum digestion and decontamination with N-acetyl-L-cysteine-sodium hydroxide for culture of mycobacteria. Am Rev Respir Dis 1963; 87: 77579. 10 Kent PT, Kubica GP. Public health mycobacteriology: a guide or the level III laboratory. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, 1985.

1188

www.thelancet.com Vol 373 April 4, 2009

Articles

11

12

13

14

15

16

17

18

Fox W, Ellard GA, Mitchison DA. Studies on the treatment of tuberculosis undertaken by the British Medical Research Council tuberculosis units, 19461986, with relevant subsequent publications. Int J Tuberc Lung Dis 1999; 3 (suppl 2): S23179. East African/British Medical Research Council. Controlled clinical trial of short-course (6-month) regimens of chemotherapy for treatment of pulmonary tuberculosis. Lancet 1972; 1: 107985. Hong Kong Chest Service/British Medical Research Council. Controlled trial of 6-month and 8-month regimens in the treatment of pulmonary tuberculosis. First report. Am Rev Respir Dis 1978; 118: 21928. Nuermberger EL, Yoshimatsu T, Tyagi S, et al. Moxioxacin-containing regimen greatly reduces time to culture conversion in murine tuberculosis. Am J Respir Crit Care Med 2004; 169: 42126. Nuermberger EL, Yoshimatsu T, Tyagi S, et al. Moxioxacin-containing regimens of reduced duration produce a stable cure in murine tuberculosis. Am J Respir Crit Care Med 2004; 170: 113134. Burman WJ, Goldberg S, Johnson JL, et al. Moxioxacin versus ethambutol in the rst 2 months of treatment for pulmonary tuberculosis. Am J Respir Crit Care Med 2006; 174: 33138. Rustomjee R, Lienhardt C, Kanyok T, et al. A phase II study of the sterilising activities of ooxacin, gatioxacin and moxioxacin in pulmonary tuberculosis. Int J Tuberc Lung Dis 2008; 12: 12838. Tuberculosis Trials Consortium. Rifapentine and isoniazid once a week versus rifampicin and isoniazid twice a week for treatment of drug-susceptible pulmonary tuberculosis in HIV-negative patients: a randomised clinical trial. Lancet 2002; 360: 52834.

19

20

21

22

23

24

25

26

Chang KC, Leung CC, Yew WW, Ho SC, Tam CM. A nested case-control study on treatment-related risk factors for early relapse of tuberculosis. Am J Respir Crit Care Med 2004; 170: 112430. Jindani A, Nunn AJ, Enarson DA. Two 8-month regimens of chemotherapy for treatment of newly diagnosed pulmonary tuberculosis: international multicentre randomised trial. Lancet 2004; 364: 124451. Park-Wyllie LY, Juurlink DN, Kopp A, et al. Outpatient gatioxacin therapy and dysglycemia in older adults. N Engl J Med 2006; 354: 135261. Pletz MW, De Roux A, Roth A, Neumann KH, Mauch H, Lode H. Early bactericidal activity of moxioxacin in treatment of pulmonary tuberculosis: a prospective, randomized study. Antimicrob Agents Chemother 2004; 48: 78082. Gosling RD, Uiso LO, Sam NE, et al. The bactericidal activity of moxioxacin in patients with pulmonary tuberculosis. Am J Respir Crit Care Med 2003; 168: 134245. Johnson J, Hadad D, Boom W, et al. Early and extended early bactericidal activity of levooxacin, gatioxacin and moxioxacin in pulmonary tuberculosis. Int J Tuberc Lung Dis 2006; 10: 60512. Gillespie SH, Gosling RD, Uiso L, Sam NE, Kanduma EG, McHugh TD. Early bactericidal activity of a moxioxacin and isoniazid combination in smear-positive pulmonary tuberculosis. J Antimicrob Chemother 2005; 56: 116971. Dorman SE, Johnson JJ, Goldberg S, et al. Moxioxacin vs isoniazid in the rst 2 months of treatment for pulmonary tuberculosis. 47th Interscience Conference on Antimicrobial Agents and Chemotherapy; Chicago, USA; Sept 1720, 2007. Abstract L-736-B.

www.thelancet.com Vol 373 April 4, 2009

1189

Articles

Coronary artery bypass surgery compared with percutaneous coronary interventions for multivessel disease: a collaborative analysis of individual patient data from ten randomised trials
Mark A Hlatky, Derek B Boothroyd, Dena M Bravata, Eric Boersma, Jean Booth, Maria M Brooks, Didier Carri, Tim C Clayton, Nicolas Danchin, Marcus Flather, Christian W Hamm, Whady A Hueb, Jan Khler, Sheryl F Kelsey, Spencer B King, Andrzej S Kosinski, Neuza Lopes, Kathryn M McDonald, Alfredo Rodriguez, Patrick Serruys, Ulrich Sigwart, Rodney H Stables, Douglas K Owens, Stuart J Pocock

Summary
Lancet 2009; 373: 119097 Published Online March 20, 2009 DOI:10.1016/S01406736(09)60552-3 See Comment page 1150 Stanford University School of Medicine, Stanford, CA, USA (Prof M A Hlatky MD, D B Boothroyd PhD, D M Bravata MD, K M McDonald MM); Department of Cardiology, Erasmus University, Rotterdam, Netherlands (Prof E Boersma PhD, Prof P Serruys MD); Royal Brompton & Harefield NHS Trust, London, UK (J Booth MS, M Flather FRCP); Department of Epidemiology, University of Pittsburgh, Pittsburgh, PA, USA (M M Brooks PhD, Prof S F Kelsey PhD); Rangueil Hospital, Toulouse, France (D Carri MD); London School of Hygiene and Tropical Medicine, London, UK (T C Clayton MSc, Prof S J Pocock PhD); AP-HP Universit Paris Descartes, Paris, France (N Danchin MD); Kerckhoff Klinik, Bad Nauheim, Germany (C W Hamm MD); Instituto do Coraco, So Paulo, Brazil (Prof W A Hueb MD, N Lopes MD); Department of Cardiology, Universitres Herzzentrum, Hamburg, Germany (J Khler MD); Saint Josephs Heart and Vascular Institute, Atlanta, GA, USA (S B King MD); Duke University School of Medicine, Durham, NC, USA (A S Kosinski PhD); Sanatario Otamendi, Buenos Aires, Argentina (A Rodriguez MD); Department of Cardiology, University of Geneva, Geneva, Switzerland

Background Coronary artery bypass graft (CABG) and percutaneous coronary intervention (PCI) are alternative treatments for multivessel coronary disease. Although the procedures have been compared in several randomised trials, their long-term eects on mortality in key clinical subgroups are uncertain. We undertook a collaborative analysis of data from randomised trials to assess whether the eects of the procedures on mortality are modied by patient characteristics. Methods We pooled individual patient data from ten randomised trials to compare the eectiveness of CABG with PCI according to patients baseline clinical characteristics. We used stratied, random eects Cox proportional hazards models to test the eect on all-cause mortality of randomised treatment assignment and its interaction with clinical characteristics. All analyses were by intention to treat. Findings Ten participating trials provided data on 7812 patients. PCI was done with balloon angioplasty in six trials and with bare-metal stents in four trials. Over a median follow-up of 59 years (IQR 50100), 575 (15%) of 3889 patients assigned to CABG died compared with 628 (16%) of 3923 patients assigned to PCI (hazard ratio [HR] 091, 95% CI 082102; p=012). In patients with diabetes (CABG, n=615; PCI, n=618), mortality was substantially lower in the CABG group than in the PCI group (HR 070, 056087); however, mortality was similar between groups in patients without diabetes (HR 098, 086112; p=0014 for interaction). Patient age modied the eect of treatment on mortality, with hazard ratios of 125 (094166) in patients younger than 55 years, 090 (075109) in patients aged 5564 years, and 082 (070097) in patients 65 years and older (p=0002 for interaction). Treatment eect was not modied by the number of diseased vessels or other baseline characteristics. Interpretation Long-term mortality is similar after CABG and PCI in most patient subgroups with multivessel coronary artery disease, so choice of treatment should depend on patient preferences for other outcomes. CABG might be a better option for patients with diabetes and patients aged 65 years or older because we found mortality to be lower in these subgroups. Funding Agency for Healthcare Research and Quality.

Introduction
Coronary artery bypass graft (CABG) and percutaneous coronary intervention (PCI) are alternative revascularisation procedures for patients with multivessel coronary artery disease. The eects of these two procedures on patient outcomes (mortality, myocardial infarction, angina symptoms, repeat procedures) over long-term follow-up have been compared in several randomised clinical trials,112 in analyses of large clinical registries,1317 and in meta-analyses of the published trial results.1820 However, the outcomes of the procedures might vary according to patient characteristics, such as the presence of diabetes or the number of diseased vessels. This possibility has been dicult to assess because no randomised trial has been large enough to provide adequate statistical power,

meta-analyses in patient subgroups have been limited by inconsistent reporting in published trials,20 and observational studies have been confounded by treatment selection biases. Pooling of individual patient data from randomised trials substantially increases the number of patients within clinical subgroups of interest and provides a more precise assessment of the eects of treatment.2124 Previous collaborations among clinical trial groups have provided information about variation in the ecacy of other cardiovascular treatments according to baseline clinical characteristics.25,26 We undertook a collaborative analysis of data from randomised trials of patients with multivessel coronary artery disease to assess whether the eects of CABG and PCI on mortality are modied by patient characteristics.
www.thelancet.com Vol 373 April 4, 2009

1190

Articles

Methods
Patients and procedures
Details of the search strategy that was used to identify relevant trials for inclusion in this collaborative analysis have been reported elsewhere.20 Briey, we searched Medline, Embase, and Cochrane databases for studies published between January, 1966, and August, 2006, by use of terms including angioplasty, coronary, and coronary artery bypass surgery. We also reviewed the reference lists of retrieved articles, conference abstracts, and the bibliographies of expert advisers. We did not limit the searches to the English language. Clinical trials that randomly assigned patients with multivessel coronary artery disease to either CABG or PCI and that reported at least 3 years of follow-up were eligible for inclusion. We excluded trials that compared either method alone with medical therapy, those that compared two forms of PCI, and those that compared two forms of CABG. All included trials were reviewed and approved by ethics committees. We identied 12 eligible trials; the principal investigators of these studies were invited to participate in this collaborative analysis.112 Investigators from ten of the trials110 provided individual patient data on a set of core
Overall (N=7812) Age <55 years 5564 years 65 years Female Diabetes Current smoker Hypertension Hypercholesterolaemia Peripheral vascular disease Unstable symptoms Previous myocardial infarction Heart failure 2185 (28%) 2933 (38%) 2688 (34%) 1831 (23%) 1233 (16%) 1665 (25%) 3503 (45%) 3386 (52%) 665 (10%) 2653 (41%) 3506 (45%) 245 (3%) 332 (28%) 420 (35%) 453 (38%) 283 (23%) 208 (17%) 323 (27%) 540 (45%) 694 (58%) 64 (5%) 451 (37%) 520 (43%) 0 (0%) 189 (17%) 338 (29%) NA 51 (5053) 580 (98%) 539 (93%) 442 (24%) 678 (37%) 709 (39%) 489 (27%) 353 (19%) 463 (25%) 896 (49%) 725 (44%) 303 (17%) 1250 (68%) 987 (55%) 161 (9%) 341 (19%) 754 (41%) 668 (37%) 104 (100110) 9 (1%) 729 (82%) ARTS1 (N=1205) BARI2 (N=1829)

clinical variables consisting of demographics (age, sex, and ethnicity), cardiac risk factors (diabetes, smoking, hypertension, and hypercholesterolaemia), clinical manifestations (stable or unstable symptoms, history of myocardial infarction, heart failure, previous PCI, previous CABG, and peripheral vascular disease), angiographic factors (abnormal left ventricular function, number of diseased vessels, and disease of the proximal left anterior descending coronary artery), randomised treatment assignment, and outcomes in follow-up (death, myocardial infarction, stroke, repeat revascularisation, last follow-up contact, and angina). We recoded data from each trial in a uniform format after resolution of data queries and checked data summaries from individual trials against the associated publications for accuracy. The primary outcome measure of this study was all-cause mortality over all available follow-up, and the principal research question was whether survival after random assignment to CABG or PCI was modied by patients baseline clinical characteristics.

(Prof U Sigwart MD); Liverpool Heart and Chest Hospital, Liverpool, UK (R H Stables DM); and VA Palo Alto Healthcare System, Palo Alto, CA, USA (Prof D K Owens MD) Correspondence to: Prof Mark A Hlatky, Stanford University School of Medicine, HRP Redwood Building, Room 150, Stanford, CA 94305-5405, USA hlatky@stanford.edu

Statistical analysis
All analyses followed the intention-to-treat principle. For descriptive analyses, we pooled individual patient
CABRI3 (N=1054) EAST4 (N=392) ERACI-II5 (N=450) GABI6 (N=323) MASS-II7 (N=408) RITA-18 (N=1011) SoS9 (N=988) Toulouse10 (N=152)

286 (27%) 443 (42%) 320 (31%) 234 (22%) 124 (12%) NA 378 (36%) 460 (44%) 72 (7%) 166 (16%) 439 (43%) 0 (0%) 138 (15%) 449 (43%) 638 (61%) 30 (2437) 0 (0%) NA

94 (24%) 143 (36%) 155 (40%) 103 (26%) 90 (23%) 79 (20%) 206 (53%) 146 (40%) NA NA 160 (41%) 13 (3%) 63 (16%) 156 (40%) 283 (72%) 82 (8282) 0 (0%) NA

124 (28%) 163 (36%) 162 (36%) 93 (21%) 78 (17%) 233 (52%) 318 (71%) 275 (61%) 103 (23%) 412 (92%) 126 (28%) 0 (0%) 88 (20%) 219 (49%) 230 (51%) 50 (5050) 221 (100%) 198 (96%)

107 (33%) 130 (40%) 86 (27%) 67 (21%) 41 (13%) 36 (11%) 136 (42%) 201 (63%) 26 (8%) 41 (13%) 150 (47%) 0 (0%) 25 (13%) 119 (38%) 92 (28%) 130 (121145) 0 (0%) 62 (39%)

131 (32%) 135 (33%) 142 (35%) 125 (31%) 115 (28%) 134 (33%) 253 (62%) 322 (79%) 0 (0%) 0 (0%) 191 (47%) 0 (0%) 13 (3%) 230 (56%) 389 (95%) 51 (5152) 157 (82%) 188 (95%)

403 (40%) 442 (44%) 166 (16%) 196 (19%) 62 (6%) 169 (17%) 265 (26%) NA NA NA 428 (43%) 0 (0%) 142 (26%) 125 (12%) 567 (56%)

253 (26%) 340 (34%) 395 (40%) 206 (21%) 142 (14%) 149 (15%) 447 (45%) 509 (52%) 66 (7%) 202 (20%) 448 (45%) 62 (6%) 153 (20%) 419 (42%) 457 (46%)

13 (9%) 39 (26%) 100 (66%) 35 (23%) 20 (13%) 79 (52%) 64 (42%) 54 (36%) 31 (20%) 131 (86%) 57 (38%) 9 (6%) 14 (9%) 44 (29%) 67 (44%) 49 (4057) 0 (0%) 42 (55%)

Abnormal left ventricular 1166 (17%) function Three-vessel disease Proximal LAD disease Follow-up (years) Stent use in PCI* IMA use in CABG 2853 (37%) 3391 (51%) 59 (50100) 1432 (37%) 2573 (83%)

100 60 (100100) (5567) 0 (0%) 364 (74%) 465 (97%) 451 (93%)

ARTS=Arterial Revascularization Therapies Study. BARI=Bypass Angioplasty Revascularization Investigation. CABRI=Coronary Angioplasty versus Bypass Revascularisation Investigation. CABG=coronary artery bypass graft. EAST=Emory Angioplasty versus Surgery Trial. ERACI=Argentine Randomised Trial of Coronary Angioplasty Versus Bypass Surgery in Multivessel Disease. GABI=German Angioplasty Bypass Surgery Investigation. IMA=internal mammary artery. LAD=left anterior descending artery. MASS=Medicine, Angioplasty, or Surgery Study. NA=not available. RITA=Randomised Intervention Treatment of Angina trial. SoS=Stent or Surgery trial. Data are n (%) or median (IQR). The number of randomised patients in each trial (N) is shown. Patients with missing data were omitted from the calculation of percentages for baseline characteristics. For trials from which data were available, between 0 and 51 (07%) patients had missing values on baseline characteristics, apart from hypercholesterolaemia (228 missing values) and left ventricular function (1066 missing values). Denitons for hypertension and hypercholesterolaemia diered among the trials. *Stent use in 3841 patients assigned to PCI who received this treatment. Eight trials provided individual patient data on IMA use, in which 3087 patients assigned to CABG received this treatment. The CABRI trial3 previously reported 81% use of IMA grafts.

Table 1: Baseline characteristics by study

www.thelancet.com Vol 373 April 4, 2009

1191

Articles

A
35 CABG PCI

B
35

30

30

25 Death or myocardial infarction (%) 0 Number of patients* CABG 3889 PCI 3923 1 3767 3798 2 3675 3709 3 4 5 Years of follow-up 3415 3431 3180 3205 2693 2658 6 1853 1828 7 1609 1576 8 1477 1452

25

20 Mortality (%)

20

15

15

10

10

0 0 3695 3725 1 3369 3419 2 3269 3310 3 4 5 Years of follow-up 3001 3023 2763 2797 2294 2267 6 1501 1491 7 1269 1253 8 1161 1150

Figure 1: Outcomes of treatment with coronary artery bypass graft or percutaneous coronary intervention CABG=coronary artery bypass graft. PCI=percutaneous coronary intervention. *Number of patients available for follow-up. Data show overall unadjusted mortality (A) and composite endpoint of death or myocardial infarction (B) after randomisation to CABG or PCI. Data on death with myocardial infarction were not available from the Emory Angioplasty versus Surgery Trial.4

5-year event rate (% [95% CI]) CABG Death Death or myocardial infarction Death or repeat revascularisation Death, myocardial infarction, or repeat revascularisation 84% (7492) 154% (142166) 99% (89109) 201% (187214) PCI 100% (90109) 167% (154179)

Hazard ratio (95% CI)*

p value

091 (082102) 097 (088106)

012 047 <00001 <00001

245% (230260) 041 (037045) 364% (348380) 052 (049057)

Event rates are unadjusted, 5-year Kaplan-Meier estimates. *Hazard ratios for coronary artery bypass graft (CABG) versus percutaneous coronary intervention (PCI) are based on the full duration of follow-up from all trials. No data were available on myocardial infarction from the Emory Angioplasty versus Surgery Trial (EAST).4 No data were available on repeat revascularisation from the Toulouse trial.10 No data were available from the EAST4 and Toulouse10 trials.

(03 years, 36 years, 69 years, and >9 years) in a stratied Cox model. Additionally, we checked for any violation of the proportional hazards assumption by testing for a correlation with follow-up time of scaled Schoenfeld residuals. We tested the eect of diabetes on mortality with and without inclusion of the trial that had previously shown an eect of diabetes on survival in patients randomised to CABG and PCI.28 We also assessed whether the method of PCI used in the trial (ie, balloon angioplasty or bare-metal coronary stents) had an eect on treatment outcome. Statistical analyses were done with SAS version 9.1 and R version 2.4.0.

Table 2: Overall clinical outcomes by treatment assignment

Role of the funding source

The sponsor of the study had no role in the study design, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had nal responsibility for the decision to submit the manuscript for publication.

data from all ten trials and created unadjusted Kaplan-Meier survival curves. For statistical analyses of mortality, we used Cox proportional hazards models stratied by trial24 that included a gamma frailty term to assess random eects across the ten contributing trials.27 We tested for interactions of assigned treatment with baseline characteristics by use of multivariable, stratied Cox models that included treatment assignment, the baseline characteristic of interest, and their interaction. We also tested the signicance of these interactions after including other baseline characteristics in the model. We undertook several analyses to test the sensitivity of results to various assumptions and model specications. Since length of follow-up varied among the trials, we tested for any dierences in the hazard ratio (HR) for CABG versus PCI as a function of follow-up time
1192

Results
The ten participating trials provided data on 7812 patients. The median age of the study population was 61 years (IQR 5367), with 389 (5%) patients aged 75 years or older (only 19 patients were aged 80 years or older). Table 1 shows the baseline characteristics of patients included in the trials. Median follow-up time in surviving patients was 59 years, and varied among trials from 30 years to 130 years (table 1). Most patients received the assigned treatment within 60 days of randomisation. Within 90 days of randomisation, 75 (2%) of 3889 patients assigned to
www.thelancet.com Vol 373 April 4, 2009

Articles

CABG died, compared with 74 (2%) of 3923 patients assigned to PCI (p=089). The composite endpoint of death or myocardial infarction within 90 days, which could be assessed in nine trials,13,510 occurred in 240 (6%) of 3695 patients in the CABG group and 201 (5%) of 3725 patients in the PCI group (p=0045). Data on stroke within 90 days of randomisation were available from seven trials:1,510 26 (1%) of 2268 patients assigned to CABG had a stroke compared with 12 (05%) of 2269 patients assigned to PCI (p=002). Overall mortality was similar between treatment groups (gure 1); 575 (15%) of 3889 patients died in the CABG group compared with 628 (16%) of 3923 patients in the PCI group (HR for mortality 091, 95% CI 082102; p=012; table 2). There was no evidence of a treatmenttime interactionie, the proportional hazards assumption was not violated. Several secondary endpoints could be assessed in most, but not all, trials (table 2). The composite endpoint of death or myocardial infarction was not signicantly
Total mortality* (n/N) CABG Age <55 years Age 5564 years Age >65 years Women Men No diabetes Diabetes Not smoking Smoking No hypertension Hypertension Normal cholesterol Hypercholesterolaemia No PVD PVD Stable symptoms Unstable symptoms No previous MI Previous MI No heart failure Heart failure Normal LV function Abnormal LV function Less than three diseased vessels Three-vessel disease No proximal LAD Proximal LAD Balloon angioplasty trials Bare-metal stent trials 107/1063 201/1477 267/1347 162/909 413/2980 432/3263 143/615 393/2558 158/816 268/2128 306/1750 236/1599 221/1667 374/2841 91/334 205/1840 262/1347 263/2123 308/1742 513/3756 59/126 375/2789 126/551 325/2386 248/1477 278/1567 249/1707 436/2356 139/1533 PCI 88/1122 220/1456 319/1341 164/922 464/3001 448/3298 179/618 440/2526 149/849 299/2167 329/1753 273/1588 247/1719 408/2872 110/331 256/1900 266/1306 286/2132 334/1764 566/3800 58/119 398/2791 151/615 371/2523 253/1376 310/1636 268/1684 481/2405 147/1518 5-year mortality (%) CABG 55% 80% 110% 96% 80% 76% 123% 79% 104% 71% 99% 90% 84% 81% 150% 82% 96% 74% 95% 75% 301% 76% 124% 77% 95% 82% 88% 85% 82% PCI 50% 94% 147% 120% 94% 81% 200% 95% 109% 87% 115% 110% 98% 91% 221% 102% 111% 93% 108% 92% 321% 91% 144% 88% 121% 102% 105% 109% 86% 05

dierent between treatment groups (gure 1). The composite outcome of death or repeat revascularisation was signicantly lower (p<00001) in patients assigned to CABG than in patients assigned to PCI (table 2). Angina at 1 year of follow-up was signicantly less frequent (p<00001) in the CABG group (439 [14%] of 3228 patients) than in the PCI group (856 [26%] of 3240 patients; dierence 13%, 95% CI 1115). Treatment eect was not modied by clinical characteristics, apart from diabetes and age (gure 2). Of the 1233 patients with diabetes, 143 (23%) of 615 patients assigned to CABG died, compared with 179 (29%) of 618 patients assigned to PCI (gure 3). By contrast, of the 6561 patients without diabetes, 432 (13%) of 3263 patients and 448 (14%) of 3298 patients died, respectively (p=0014 for interaction). The interaction of diabetes with treatment remained after adjustment for age, sex, smoking, hypertension, history of myocardial infarction, heart failure, and three-vessel disease (p=0008), and also after exclusion of patients
Hazard ratio (95% CI)* p value

125 (094166) 090 (075109) 082 (070097) 102 (082127) 088 (077100) 098 (086112) 070 (056087) 087 (076100) 111 (089139) 090 (076106) 093 (079108) 084 (071100) 093 (077111) 092 (080106) 078 (059103) 083 (069099) 095 (080112) 092 (078109) 091 (078107) 091 (080102) 101 (070146) 092 (080106) 093 (073118) 091 (078106) 091 (077109) 092 (079109) 090 (075107) 091 (080103) 094 (074118) 08 Favours CABG 10 125 Favours PCI 20

0002

025 0014 0073 073 046 033 030 092 046 087 098 077 019

Figure 2: Subgroup analyses for mortality after treatment with coronary artery bypass graft or percutaneous coronary intervention CABG=coronary artery bypass graft. LAD=left anterior descending artery. LV=left ventricular. MI=myocardial infarction. PCI=percutaneous coronary intervention. PVD=peripheral vascular disease. The vertical line indicates a hazard ratio of 10, equivalent to no dierence between treatment groups. *Based on on the full duration of follow-up in all trials. Pooled unadjusted 5-year Kaplan-Meier survival rates. p value for the treatment by covariate interaction. The analysis that compares patients enrolled in balloon angioplasty trials24,6,8,10 and bare-metal stent trials1,5,7,9 is pooled and not stratied by study.

www.thelancet.com Vol 373 April 4, 2009

1193

Articles

A
35 CABG no diabetes CABG diabetes PCI no diabetes PCI diabetes

30

25

Mortality (%)

20

15

10

0 0 Number of patients* CABG no diabetes 3263 CABG diabetes 615 PCI no diabetes 3298 PCI diabetes 618 1 2 3 4 5 Years of follow-up 2877 532 2918 508 2677 498 2725 475 2267 421 2281 373 6 1592 257 1608 218 7 1380 225 1393 179 8 1274 200 1288 160 0 2529 435 2556 445 1 2457 420 2493 421 2 2382 410 2432 408 3 4 5 Years of follow-up 2179 371 2215 369 1992 344 2031 344 1598 278 1606 258 6 940 120 946 110 7 747 91 750 81 8 655 73 655 66

3169 3089 587 575 3217 3148 574 555

Figure 3: Mortality in patients assigned to coronary artery bypass graft or percutaneous coronary intervention by diabetes status CABG=coronary artery bypass graft. PCI=percutaneous coronary intervention. *Number of patients available for follow-up. Data show overall unadjusted mortality rates for patients with diabetes and without diabetes. Panel A includes patients from all ten trials. Panel B excludes patients from the Bypass Angioplasty Revascularization Investigation trial.2

35

30

CABG <55 years CABG 5564 years CABG 65 years PCI <55 years PCI 5564 years PCI 65 years

25

20

15

10

0 0 Number of patients* CABG <55 years CABG 5564 years CABG 65 years PCI <55 years PCI 5564 years PCI 65 years 1063 1477 1347 1122 1456 1341 1 1039 1434 1292 1104 1404 1286 2 1016 1400 1257 1087 1371 1247 3 4 5 Years of follow-up 891 1206 1082 955 1177 1072 774 1017 902 817 989 852 6 556 711 586 561 705 562 7 507 634 468 500 614 462 8 468 585 424 466 578 408

938 1308 1167 1013 1262 1154

Figure 4: Mortality in patients assigned to coronary artery bypass graft or percutaneous coronary intervention by age CABG=coronary artery bypass graft. PCI=percutaneous coronary intervention. *Number of patients available for follow-up. Data show overall unadjusted mortality rates for patients aged less than 55 years, 5564 years, and 65 years or older.

enrolled in the Bypass Angioplasty Revascularization Investigation (BARI) trial28 (HR 068, 047095, in patients with diabetes; HR 101, 085120, in patients without diabetes; p=0048 for interaction; gure 3). Patient age had a graded eect on mortality after CABG or PCI (p=0002 for interaction with age as a continuous variable; gure 2 and gure 4). 107 (10%) of 1063 patients younger than 55 years who were assigned to CABG died compared with 88 (8%) of 1122 patients assigned to PCI. 201 (14%) of 1477 patients aged 5564 years in the CABG group died compared with 220 (15%) of 1456 patients in the PCI group. In patients aged 65 years and older, mortality was 20% (267 of 1347 patients) for CABG and 24% (319 of 1341 patients) for PCI. The interaction between age and treatment eect remained after adjustment for sex, diabetes, smoking, hypertension, history of myocardial infarction, heart failure, and three-vessel disease (p=0002). In the six earliest trials,24,6,8,10 PCI was done with balloon angioplasty, whereas in the four more recent trials, the procedure was done with bare-metal stents.1,5,7,9 Most baseline clinical characteristics diered signicantly (p<00001) between patients in the two types of trial (table 3). There was no signicant dierence in survival between CABG and PCI groups when assessed by bare-metal stent or balloon angioplasty (gure 2). In the six balloon angioplasty trials, 436 (19%) of 2356 patients died in the CABG group compared with 481 (20%) of 2405 patients in the PCI group, whereas in
www.thelancet.com Vol 373 April 4, 2009

1194

Mortality (%)

Articles

the bare-metal stent trials 139 (9%) of 1533 patients and 147 (10%) of 1518 patients died, respectively (gure 2). In a multivariable analysis of pooled data that adjusted for baseline patient characteristics and restricted length of follow-up to a maximum of 5 years, there was no signicant eect of trial use of bare-metal stents on the treatment comparison of CABG and PCI (p=019 for interaction). The interactions of diabetes and age with treatment assignment that were present in the overall population were evident in both balloon angioplasty and bare-metal stent trials (data not shown).

Balloon angioplasty Bare-metal stent trials1,5,7,9 trials24,6,8,10 (N=4761) (N=3051) Age <55 years 5564 years 65 years Female Diabetes Current smoker Hypertension Hypercholesterolaemia Peripheral vascular disease Unstable symptoms Previous myocardial infarction Heart failure Abnormal left ventricular function Three-vessel disease Proximal LAD disease 1345 (28%) 1875 (39%) 1536 (32%) 1124 (24%) 690 (15%) 826 (22%) 1945 (41%) 1586 (45%) 432 (13%) 1588 (48%) 2221 (47%) 183 (4%) 723 (18%) 1647 (35%) 2315 (49%) 840 (28%) 1058 (35%) 1152 (38%) 707 (23%) 543 (18%) 839 (28%) 1558 (51%) 1800 (59%) 233 (8%) 1065 (35%) 1285 (42%) 62 (2%) 443 (16%) 1206 (40%) 1076 (58%)

p value

<00001 066 00001 <00001 <00001 <00001 <00001 <00001 <00001 <00001 004 <00001 <00001

Discussion
Randomised clinical trials provide the reference standard for comparing the eectiveness of treatments for a given clinical condition. The eectiveness of treatments might vary among patients included in randomised trials, but this possibility cannot be tested adequately in a single study because of limited statistical power. Combining individual patient data from several randomised trials helps to overcome this limitation by increasing the number of patients available for analysis in clinical subgroups, thus enhancing statistical power. Our combined analysis of individual patient data from ten randomised trials suggests that diabetes and age modify the eect of CABG compared with PCI on the survival of patients with multivessel coronary disease. Treatment eect was not altered by other patient characteristics, including the number of diseased coronary vessels, despite observational data strongly suggesting that this factor would modify the eectiveness of coronary revascularisation.13,14,16 The pooled data provide more precise estimates of the overall eect of CABG and PCI on long-term survival, both overall and within clinical subgroups. The BARI trial28 was the rst to report that patients with diabetes had substantially better survival after CABG than after PCI. This result was not universally accepted, since analyses of large clinical registries did not conrm this eect;29,30 similarly, other, smaller randomised clinical trials were unable to replicate the BARI trial ndings.4,6,8,31 Our analysis is based on pooled data from 1233 randomised patients with diabetes and provides strong evidence that survival is substantially higher after CABG than PCI for the treatment of multivessel disease. This nding is not a result of the inclusion of the BARI trial,2 since a signicant interaction of diabetes with treatment assignment remained after exclusion of that trial. Nor is our result explained by the adverse clinical risk prole of patients with diabetes, because it remained signicant after adjustment for other baseline clinical characteristics. Despite the strength of our nding, it is important to note that coronary revascularisation and background medical treatment have continued to advance since the trials in this study were done. Further evidence in this long-running debate will be provided by the results of current trials of the procedures in patients with diabetes.32,33
www.thelancet.com Vol 373 April 4, 2009

LAD=left anterior descending artery. Data are n (%). Patients with missing data were omitted from the calculations of percentages for baseline characteristics.

Table 3: Clinical characteristics before randomisation to treatment by percutaneous coronary intervention method used in the trial

Our nding that patient age modies the relative eectiveness of CABG and PCI on survival has not previously been reported by individual randomised trials. The interaction of age with assigned treatment might be mediated by the more favourable clinical characteristics in younger patients; however, we found that the eect persisted after multivariable adjustment for such characteristics. One potential interpretation of this nding is that younger patients might benet more from initial PCI than from CABG because the latter treatment could be done at a more appropriate time in the course of their disease. Another potential interpretation is that older age might be a marker for more severe disease that was otherwise unmeasured and that might respond better to CABG. It is important to emphasise that few patients in this study were 75 years or older, and the older patients randomised in these trials might have been more highly selected. Observational comparisons of CABG with PCI suggest a strong relation between the extent of coronary disease and the relative eectiveness of these procedures on survival.13,14,16 In particular, clinical registry studies have reported that patients with the least extensive coronary disease have better survival after PCI, whereas patients with the most extensive disease have better survival after CABG.13 Contrary to these observational data and to our previous hypothesis, we found no signicant interaction between the number of diseased vessels and treatment eect. An association might have been found if we had been able to analyse a more detailed measure of extent of disease, such as the Duke13 or SYNTAX34 scores. However, a count of diseased vessels was the only
1195

Articles

measure available from all ten trials. The extent of disease in patients eligible for randomisation might also have fallen into a narrow range in which CABG and PCI yield equivalent results.13 Additionally, the results of observational studies might represent the residual eects of selection bias rather than a true variation in clinical eectiveness, since the extent of coronary disease is the strongest clinical factor aecting the choice between CABG and PCI for coronary revascularisation.17 The techniques of the procedures investigated here continue to be rened over time, and coronary stents in particular have been widely adopted for PCI. Six of the trials included in this analysis were done before the introduction of coronary stents,24,6,8,10 whereas the remaining four studies1,5,7,9 were done after bare-metal stents became available. We attempted to assess whether the results of the earlier trials diered from those of the subsequent trials. This analysis was dicult because stent use was completely confounded with patient enrolment in specic trials. There were also many other dierences between these trials, including important dierences in baseline clinical characteristics (table 3). We found that the eect of CABG compared with PCI on survival did not dier between balloon angioplasty and bare-metal stent trials. This result is consistent with the ndings from meta-analyses of randomised trials that showed no signicant dierence in survival, despite signicant reductions in the rate of repeat revascularisation procedures, between balloon angioplasty PCI and bare-metal stents,35 or between bare-metal stents and drug-eluting stents.36 Our study shows that the pooling of individual patient data from randomised trials to assess treatment has advantages over the more common technique of meta-analysis of published aggregate data. Most clinical trials do not publish results in key subgroups of interest,20 and even when they do, the data are typically presented in dierent ways and are dicult to combine in a meta-analysis. Pooling of individual patient data overcomes these limitations, and also allows use of more sensitive statistical methods, including analysis of survival times, use of multivariable models, and tests for treatment-by-covariate interactions. However, this technique poses logistical challenges and requires collaboration among trial groups and support from funding agencies; thus, it has not been used as often as meta-analysis of published data. Our experience suggests that collaborative analysis could be used more often, especially to assess subgroup eects that are dicult to address in one trial. Our study has several limitations. We were not able to obtain data from two smaller trials of CABG and PCI that enrolled 359 patients with multivessel disease,11,12 but we did analyse data from 95% of all randomised patients, and believe our results would be unlikely to change if these smaller trials were included. We have no data on
1196

concomitant drug treatment or on control of coronary risk factors during follow-up. Our analysis shares the underlying limitations of the ten participating trials, which excluded some patients of interest (eg, those with previous CABG or PCI), and did not have adequate representation of others (eg, patients aged 75 years and older or patients with reduced left ventricular function). The participating studies selected patients in whom either treatment would be technically feasible and for whom either would be a reasonable clinical option. Consequently, patients with extensive three-vessel disease or left main coronary artery disease were generally excluded because CABG would be the most appropriate treatment, and patients with limited single-vessel disease were excluded because PCI would be most appropriate. Therefore, our ndings should not be extrapolated to all patients with coronary disease; they apply only to patients for whom either CABG or PCI is a reasonable therapeutic option and to patients similar to those enrolled in the contributing trials. None of the ten trials included in this study used drug-eluting stents for PCI. Although clinical trials have shown equivalent rates of mortality and myocardial infarction after randomisation to either bare-metal stents or drug-eluting stents,36 trials that compare CABG with PCI by use of drug-eluting stents are still in progress.33,37 The recently reported 1-year follow-up from the SYNTAX trial,37 which showed no signicant dierence in the combined endpoint of death, myocardial infarction, or stroke between patients randomly assigned to CABG or to PCI with drug-eluting stents, are generally consistent with the results of our combined analysis. Thus, pooled data from ten long-term randomised trials of patients with multivessel coronary disease suitable for either CABG or PCI suggest that patients with diabetes, and older patients, might have a signicant survival advantage if treated with CABG.
Contributors MAH, DBB, DMB, KMM, and DKO designed the collaborative analysis. The individual patient data were collected and prepared for analysis by the investigators representing the ten participating trials: EB, PS (ARTS);1 MMB, SFK (BARI);2 ND (CABRI);3 SBK, ASK (EAST);4 AR (ERACI-II);5 CWH, JK (GABI);6 WAH, NL (MASS-II);7 TCC, SJP (RITA-I);8 JB, TCC, MF, US, RHS (SoS);9 and DC (Toulouse).10 DBB and MAH participated in data analysis. MAH drafted the manuscript. All authors reviewed and revised the manuscript, and approved the nal version. Conict of interest statement The authors declare research support from Boston Scientic (SJP, RHS, TCC), Cordis (RHS, MF), Eli Lilly (MF), Medtronic (RHS, MF), Boehringer Ingleheim (RHS), Pzer (ND), Sano-Aventis (MF), and Servier (ND); serving as consultants or on advisory boards to Boston Scientic (RHS), Medtronic (RHS), Cordis (RHS), Abbot (RHS), Eli Lilly (RHS), and Sano-Aventis (RHS); and receiving speaking fees from AstraZeneca (ND), Boston Scientic (RHS), Medtronic (RHS), Cordis (RHS, ND), Eli-Lilly (RHS, ND), Sano-Aventis (RHS, ND), Bristol Myers Squibb (RHS, ND), Boehringer Ingleheim (RHS, ND), GlaxoSmithKline (ND), Novartis (ND), Pzer (ND), Servier (ND), MSD (ND), and Takeda (ND). All other authors declare that they have no conict of interest.

www.thelancet.com Vol 373 April 4, 2009

Articles

Acknowledgments This manuscript is derived from work supported under a contract with the Agency for Healthcare Research and Quality, Rockville, MD, USA (#290-02-0017); some of the contributing trials were initially conducted with industrial support. We thank Artyom Sedrakyan for his support of this project. Our late colleagues, Katherine Detre of the BARI trial2 and Anthony Rickards of the CABRI trial,3 were important contributors to the investigation of CABG and PCI. Disclaimer: This article has not yet been approved by the funding agency. The authors of this article are responsible for its contents. No statement in this article should be construed as an ocial position of the Agency for Healthcare Research and Quality or the US Department of Health and Human Services. References 1 Serruys PW, Ong ATL, van Herwerden LA, et al. Five-year outcomes after coronary stenting versus bypass surgery for the treatment of multivessel disease. The nal analysis of the Arterial Revascularization Therapies (ARTS) randomized trial. J Am Coll Cardiol 2005; 46: 57581. 2 BARI Investigators. The nal 10-year follow-up results from the BARI randomized trial. J Am Coll Cardiol 2007; 49: 160006. 3 CABRI Trial Participants. First-year results of CABRI (Coronary Angioplasty versus Bypass Revascularisation Investigation). Lancet 1995; 346: 117984. 4 King SB, Kosinski AS, Guyton RA, Lembo NJ, Weintraub WS. Eight-year mortality in the Emory Angioplasty Versus Surgery Trial (EAST). J Am Coll Cardiol 2000; 35: 111621. 5 Rodriguez AE, Baldi J, Pereira CF, et al. Five-year follow-up of the Argentine randomized trial of coronary angioplasty with stenting versus coronary bypass surgery in patients with multiple vessel disease (ERACI II). J Am Coll Cardiol 2005; 46: 58288. 6 Kaehler J, Koester R, Billmann W, et al. 13-year follow-up of the German angioplasty bypass surgery investigation. Eur Heart J 2005; 26: 214853. 7 Hueb W, Lopes NH, Gersh BJ, et al. Five-year follow-up of the Medicine, Angioplasty, or Surgery Study (MASS II). A randomized controlled clinical trial of 3 therapeutic strategies for multivessel coronary artery disease. Circulation 2007; 115: 108289. 8 Henderson RA, Pocock SJ, Sharp SJ, et al. Long-term results of RITA-1 trial: clinical and cost comparisons of coronary angioplasty and coronary-artery bypass grafting. Lancet 1998; 352: 141925. 9 Booth J, Clayton T, Pepper J, et al. Randomized, controlled trial of coronary artery bypass surgery versus percutaneous coronary intervention in patients with multivessel coronary artery disease. Six-year follow-up from the Stent or Surgery trial (SoS). Circulation 2008; 118: 38188. 10 Carri D, Elbaz M, Puel J, et al. Five-year outcome after coronary angioplasty versus bypass surgery in multivessel coronary artery disease: results from the French Monocentric Study. Circulation 1997; 96 (suppl II): 16. 11 Morrison DA, Sethi G, Sacks J, et al. Percutaneous coronary intervention versus coronary artery bypass graft surgery for patients with medically refractory myocardial ischemia and risk factors for adverse outcomes with bypass: a multicenter randomized trial. J Am Coll Cardiol 2001; 38: 14349. 12 Rodriguez A, Mele E, Peyregne E, et al. Three-year follow-up of the Argentine randomized trial of percutaneous transluminal coronary angioplasty versus coronary artery bypass surgery in multivessel disease (ERACI). J Am Coll Cardiol 1996; 27: 117884. 13 Mark DB, Nelson CL, Cali RM, et al. Continuing evolution of therapy for coronary artery disease. Initial results from the era of coronary angioplasty. Circulation 1994; 89: 201525. 14 Smith PK, Cali RM, Tuttle RH, et al. Selection of surgical or percutaneous coronary intervention provides dierential longevity benet. Ann Thorac Surg 2006; 82: 142029. 15 Hannan EL, Racz MJ, Walford G, et al. Long-term outcomes of coronary-artery bypass grafting versus stent implantation. N Engl J Med 2005; 352: 217483. 16 Malenka DJ, Leavitt BJ, Hearne MJ, et al. Comparing long-term survival of patients with multivessel coronary disease after CABG or PCI. Analysis of BARI-like patients in Northern New England. Circulation 2005; 112 (suppl I): 37176.

17

18

19

20

21

22

23

24

25

26

27 28

29

30

31

32

33

34

35 36

37

Brener SJ, Lytle BW, Casserly IP, Schneider JP, Topol EJ, Lauer MS. Propensity analysis of long-term survival after surgical or percutaneous revascularization in patients with multivessel coronary artery disease and high-risk features. Circulation 2004; 109: 229095. Pocock SJ, Henderson RA, Rickards AF, et al. Meta-analysis of randomised trials comparing coronary angioplasty with bypass surgery. Lancet 1995; 346: 118489. Homan SN, TenBrook JA, Wolf MP, Pauker SG, Salem DN, Wong JB. A meta-analysis of randomized controlled trials comparing coronary artery bypass graft with percutaneous transluminal coronary angioplasty: one- to eight-year outcomes. J Am Coll Cardiol 2003; 41: 1293304. Bravata DM, Gienger AL, McDonald KM, et al. Systematic review: the comparative eectiveness of percutaneous coronary interventions and coronary artery bypass graft surgery. Ann Intern Med 2007; 147: 70316. Stewart LA, Parmar MKB. Meta-analysis of the literature or of individual patient data: is there a dierence? Lancet 1993; 341: 41822. Thompson SG, Higgins JPT. Can meta-analysis help target interventions at individuals most likely to benet? Lancet 2005; 365: 34146. Stewart LA, Clarke MJ. Practical methodology of meta-analyses (overviews) using updated individual patient data. Stat Med 1995; 14: 205779. Simmonds MC, Higgins JPT, Stewart LA, Tierney JF, Clarke MJ, Thompson SG. Meta-analysis of individual patient data from randomized trials: a review of methods used in practice. Clin Trials 2005; 2: 20917. Fibrinolytic Therapy Trialists (FTT) Collaborative Group. Indications for brinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity results from all randomised trials of more than 1000 patients. Lancet 1994; 343: 31122. Flather MD, Yusuf S, Kber L, et al. Long-term ACE-inhibitor therapy in patients with heart failure or left-ventricular dysfunction: a systematic overview of data from individual patients. Lancet 2000; 355: 157581. Therneau TM, Grambsch PM, Pankratz VS. Penalized survival models and frailty. J Comput Graph Stat 2003; 12: 15675. Bypass Angioplasty Revascularization Investigation (BARI) Investigators. Comparison of coronary bypass surgery with angioplasty in patients with multivessel disease. N Engl J Med 1996; 335: 21725. Barsness GW, Peterson ED, Ohman EM, et al. Relationship between diabetes mellitus and long-term survival after coronary bypass and angioplasty. Circulation 1997; 96: 255156. Niles NW, McGrath PD, Malenka D, et al. Survival of patients with diabetes and multivessel coronary artery disease after surgical or percutaneous coronary revascularization: results of a large regional prospective study. J Am Coll Cardiol 2001; 37: 100815. Kurbaan AS, Bowker TJ, Ilsley CD, Sigwart U, Rickards AF. Dierence in the mortality of the CABRI diabetic and nondiabetic populations and its relation to coronary artery disease and the revascularization mode. Am J Cardiol 2001; 87: 94750. Kapur A, Malik IS, Bagger JP, et al. The coronary artery revascularisation in diabetes (CARDia) trial: Background, aims, and design. Am Heart J 2005; 149: 1319. Farkouh ME, Dangas G, Leon MB, et al. Design of the future revascularization evaluation in patients with diabetes mellitus: optimal management of multivessel disease (FREEDOM) trial. Am Heart J 2008; 155: 21523. Ong ATL, Serruys PW, Mohr FW, et al. The synergy between percutaneous coronary intervention with taxus and cardiac surgery (SYNTAX) study: design, rationale, and run-in phase. Am Heart J 2006; 151: 1194204. Brophy JM, Belisle P, Joseph L. Evidence for use of coronary stents. A hierarchical Bayesian meta-analysis. Ann Intern Med 2003; 138: 77786. Babapulle MN, Joseph L, Blisle P, Brophy JM, Eisenberg MJ. A hierarchical Bayesian meta-analysis of randomised clinical trials of drug-eluting stents. Lancet 2004; 364: 58391. Serruys PW, Morice MC, Kappetein AP, et al. Percutaneous coronary intervention versus coronary-artery bypass grafting for severe coronary artery disease. N Engl J Med 2009; 360: 96172.

www.thelancet.com Vol 373 April 4, 2009

1197

New Drug Class

Novel opioid antagonists for opioid-induced bowel dysfunction and postoperative ileus
Gerhild Becker, Hubert E Blum
Lancet 2009; 373: 1198206 Published Online February 13, 2009 DOI:10.1016/S01406736(09)60139-2 Department of Palliative Care (G Becker MD) and Department of Internal Medicine II (Prof H E Blum MD), University Hospital Freiburg, Freiburg, Germany Correspondence to: Dr Gerhild Becker, Department of Palliative Care, University Hospital Freiburg, Robert-Koch-Str 3, D-79106 Freiburg, Germany gerhild.becker@ uniklinik-freiburg.de

Peripherally acting -opioid receptor antagonists methylnaltrexone and alvimopan are a new class of drugs designed to reverse opioid-induced side-eects on the gastrointestinal system without compromising pain relief. This article gives an overview of the pharmacology, the ecacy, and adverse eects of these drugs. Both compounds seem to be generally well tolerated and eective for the treatment of opioid-related bowel dysfunction and postoperative ileus. Methylnaltrexone recently received approval by the US Food and Drug Administration (FDA) and the European Medicines Agency for treatment of opioid-related bowel dysfunction in patients with advanced illness. Alvimopan was recently approved by the FDA for treatment of postoperative ileus, but the use of the drug is restricted to inpatients because it has been associated with an increased rate of myocardial infarction. Further research should assess the eectiveness and safety of these drugs in clinical practice.

Introduction
The peripherally acting -receptor antagonists methylnaltrexone and alvimopan are a new class of drugs designed to reverse opioid-induced side-eects on the gastrointestinal tract without compromising pain relief. Opioids are a mainstay in the treatment of acute and chronic painin 2005, opioids were prescribed about 365 million times worldwide.1 Although opioids are very eective for pain relief from cancer and non-malignant diseases, their use is often limited by side-eects. The most common side-eect is constipation.2,3 Schug and colleagues4 showed that in patients with cancer, constipation and vomiting were the most common adverse eects associated with opioid therapy. However, opioid therapy also aects bowel function by causing opioid-induced bowel dysfunction. This largely under-recognised condition is a symptom complex

Search strategy and selection criteria We searched Medline, including Medline in Process and other non-indexed citations (1950November, 2008), Embase (1980November, 2008), Cochrane Database of Systematic Reviews (quarter 4, 2008), Cochrane Central Register of Controlled Trials (quarter 4, 2008), Database of Abstracts of Reviews of Eects (quarter 4, 2008), and BIOSIS Previews (1969November, 2008), with the OVID interface, using search terms including alvimopan, methylnaltrexone, ADL 8-2698, LY246736, constipation, intestinal obstruction, opioid-induced bowel disease, and postoperative ileus. We also searched PubMed (1966November, 2008), ACP-Journal Club (1991November, 2008), and CINAHL (1982November, 2008). We identied articles about peripherally acting opioid antagonists and opioid-induced constipation and postoperative ileus, before limiting the search results to clinical trials in human beings. No language restrictions were placed on the searches. Publications were largely selected from the past 4 years but also included relevant articles from a systematic review (19662005, reference 30). We scanned reference lists of studies from ISIs Database Web of Science (1945November, 2008), and did internet searches with the terms listed above using the science-specic search engines Scirus and Google Scholar. Further trials were identied from Current Controlled Trials, WHO International Clinical Trials Registry Platform , National Institutes of Health Randomized Trial Records, and Pharmaceutical Industry Clinical Trials Database. The date of the last search was Nov 27, 2008, and we last repeated some of the search on Jan 9, 2009, especially to check for relevant research articles.

characterised by accumulation of gas and secretions, and retention of bowel content, leading to hard stool, incomplete evacuation, bloating, pain, nausea, and vomiting.2,5 Opioid-induced bowel dysfuntion can occur immediately after the rst dose and persist for the duration of therapy. Unlike many of the other side-eects, patients rarely develop tolerance to the constipating eects of opioids.6,7 Traditional treatment includes the use of stool softeners, stimulants, and osmotic agents.68 Presently no treatment exists for management of the condition beyond laxatives that specically target the pharmacological action of opioids. Postoperative ileus is a block of coordinated bowel motility after surgery.810 The pathophysiology is multifactorial and incompletely understood.911 Major mechanisms include surgical stress from physical manipulation of the bowel, secretion of inammatory mediators, endogenous opioids in the gastrointestinal tract, and changes to uid balance, and hormone and electrolyte concentrations.1214 Opioids given for postoperative pain importantly contribute to the condition by inhibition of propulsive gastrointestinal motility.15,16 Furthermore, the evidence for immunomodulatory eects of opioids is increasing1720eg, the extended phase of postoperative ileus is caused by an enteric inammatory response and recruitment of leucocytes to the muscularis of the bowel wall.16 Leucocytes produce the main inhibitory neurotransmitter of the gastrointestinal tract, nitric oxide, through the activity of the inducible isoform of nitric oxide synthase (iNOS).21 Opioids potentiate iNOS induction and nitric oxide release from phagocytes;18 therefore, blockage of this opioid-mediated response might reverse postoperative ileus.16 The condition can cause pain and discomfort, prevent enteral nutrition, and thereby extend time in hospital.9 Consequently, eective management is highly desirable. Treatment for postoperative ileus consists of intravenous hydration, use of a nasogastric tube, metoclopramide, neostigmine, and the o-label use of various drugseg, erythromycin or somatostatin.13,22,23 According to a systematic review,22 erythromycin has
www.thelancet.com Vol 373 April 4, 2009

1198

New Drug Class

CNS Opioids bind to central -opioid receptors to provide analgesia Endogenous opioids Met-enkephalin, leu-enkephalin, -endorphin, and dynorphin (localised in neurons of myenteric and submucosal plexus, and in endocrine cells of mucosa) Exogenous opioids (eg, morphine and codeine)

Opioid Opioids continue activating -opioid receptors No reversal of analgesia No causation of opioid withdrawal

-opioid receptor Bloodbrain barrier

Peripheral opioid antagonists do not cross bloodbrain barrier

Opioid

Peripheral opioid antagonist No adverse gastrointestinal eects Peripheral opioid antagonists dislodge opioid from -opioid receptors in gastrointestinal tract

Gastrointestinal tract Opioids bind to peripheral -opioid receptors, which inhibits bowel function

-opioid receptor

Figure 1: Action of opioids and peripheral -opioid receptor antagonists in the CNS and gastrointestinal tract

consistent absence of eect, evidence is insucient for cholecystokinin-like drugs (cisapride, dopamine-antagonists, propranolol, and vasopressin), and intravenous lidocaine and neostigmine might show an eect, but more evidence on clinically relevant outcomes is needed. Two further drugs are being investigated: atilmotin,24 a receptor agonist of the endogenous hormone motilin, which increases upper gastrointestinal tract motility; and lubiprostone,25,26 a chloride-channel activator that acts locally in the gut to increase intestinal uid secretion, and has been approved by the US Food and Drug Administration (FDA) for idiopathic chronic obstipation. Until now, however, no specic treatment has existed to prevent postoperative ileus or to shorten its course.9,27 Postoperative ileus and opioid-induced bowel dysfuntion are serious conditions that impose considerable expense on the health-care system. Retrospective review of the use of ICD-9 codes to assess postoperative ileus in more than 800 000 US surgical patients, showed that the average frequency of the condition for all types of surgery was 45%.28 The average length of hospital stay was 93 days for patients with postoperative ileus compared with 53 days for those without. For patients with the condition, mean total hospital costs increased substantially by US$6300. Application of the occurrence of postoperative ileus and its related costs to the 425 million inpatient surgeries, which took place in 2002, translates to more than $11 billion of increased hospital costs.29 Therefore new drugs such as methylnaltrexone and alvimopan hold promise for resolving a clinically relevant problem.

opioids act predominantly via the receptor.7,16 Such receptors are present in the CNS and peripheral nervous system (in the CNS they are the primary receptors in pain management31,32) and enteric receptors seem to be the principal mediator of opioid eects on the gastrointestinal tract (panel).34,35 Ideally, the adverse gastrointestinal eects of endogeneous and exogeneous opioids on opiate-related bowel dysfunction and postoperative ileus would be specically inhibited, without reversal of the analgesic eect of opioids in the CNS. Opioid antagonists with limited systemic absorption were the rst agents to be used for this purposeeg, -opioid receptor-specic antagonists naloxone, nalmefene, and naltrexone. These
Panel: Opioid eects in the gastrointestinal tract Inhibition of enteric nerve activity Suppression of enteric nerve excitability Presynaptic and postsynaptic inhibition of transmission in excitatory motor, inhibitory motor, and secretomotor pathways Inhibition of release of substance P, nitric oxide, and acetylcholine Inhibition of propulsive motor activity Increase of muscle tone Inhibition of distension-induced peristalsis Induction of non-propulsive motility patterns Disturbance of migrating myoelectrical complex Inhibition of gastric emptying Delay of gastrointestinal transit Inhibition of secretory activity Inhibition of ion and uid secretion Immune activity Alteration of immune cell function
Adapted from Holzer et al.33

Mechanism of action
Receptor-binding studies and molecular cloning techniques have identied three main classes of opioid receptors: , , and , with several subtypes in each class.30 These opioid receptors are stimulated by both endogenous (enkephalins, endorphins, and dynorphins) and exogenous (morphine and codeine) agonists (gure 1).16,30 Exogenous
www.thelancet.com Vol 373 April 4, 2009

1199

New Drug Class

CH3 N+ H

HO Br and epimer at N+

HO

O H

For the Martindale website see http://www.medicinescomplete. com/mc/

Figure 2: Chemical structure of methylnaltrexone bromide Structure taken from Martindale.

tertiary antagonists readily crossed the bloodbrain barrier and failed to consistently restore gastrointestinal function without reversing analgesia or inducing opioid withdrawal.30,36 Subsequently, peripheral opioid antagonists that are not systemically absorbed and do not penetrate the bloodbrain barrier were developed, and we will discuss the novel peripheral antagonists methylnaltrexone and alvimopan.

obtained exclusive worldwide rights and they formed a collaboration with Wyeth Pharmaceuticals, Madison, NJ, USA, in 2005 to develop and market the compound (Relistor). Methylnaltrexone is a quaternary derivative of naltrexone, which results in a distinct pharmacological prole (table 1). Addition of a methyl group to the nitrogen increases polarity and reduces lipid solubility, which prevents the drug crossing the bloodbrain barrier.4648 Demethylation might allow the tertiary compound to cross this barrier more readily. Methylnaltrexones metabolism is species-dependent: in rats and mice the drug is demethylated over time, whereas demethylation is negligible in dogs and primates.49,50 Methylnaltrexone antagonises opioid binding at the -opioid receptor (median inhibitory concentration IC50=70 nmol/L), but has lower anity to the receptor (IC50=575 nmol/L), and negligible anity to the receptor.48 Toxicity studies in rats showed a high median lethal dose of more than 100 mg/kg; in primates the compound was given in doses of up to 50 mg/kg without mortality.46

Ecacy
Preclinical and early clinical studies have shown that methylnaltrexone antagonises opioid-mediated inhibition of gastrointestinal function in the periphery without antagonising CNS-mediated eects such as analgesia or pupillary constriction.37,48,51 The peripheral nature of opioid-receptor antagonism in animal models has been conrmed in healthy volunteers5153 and members of methadone programmes to mimic the condition of patients on opioids (table 2).54,55 Methylnaltrexone is ecacious for treatment of opioid-induced bowel dysfunction when it is given intravenously, subcutaenously, orally, and orally with an enteric-coated formulation.5155,58,59 McNicols60 systematic review of four studies showed that, on average, gastrointestinal transit time in patients given methylnaltrexone was reduced by 52 min (95% CI 73 to 32 min) compared with placebo.51,53,54,59 Methylnaltrexone reduced the mean transit time to 93110 min from 140163 min for palcebo. Phase I and II studies have been limited by small sample size and inherent weaknesses,30 but the methodology of most studies is good, and together with preclinical data, these studies show proof of concept. Subsequently, several randomised double-blind placebocontrolled phase II and III trials have shown ecacy of methylnaltrexone in patients with advanced illnesses and opioid-induced bowel dysfunction.6164 One phase III trial showed that the drug was at least three times more ecacious than placebo in producing laxation within 4 h of the initial dose, which seemed to be undiminished throughout the 3-month open-label extension.56 Methylnaltrexone has also been investigated for prevention of postoperative ileus. A phase II trial showed treatment accelerated time to rst bowel movement and discharge eligibility,57 and phase III trials are continuing.65
www.thelancet.com Vol 373 April 4, 2009

Methylnaltrexone
Pharmacology
Methylnaltrexone (gure 2) was developed at the University of Chicago, Chicago, IL, USA, and out-licensed to UR Labs in 1985. In 2001, Progenics Pharmaceuticals

Methylnaltrexone Chemistry Receptor antagonist

Alvimopan

Derivative of naltrexone; molecular weight Synthetic; molecular weight 4601 Da; 4363 Da; positively charged zwitterion (dipolar) Peripheral -opioid receptor antagonist; 8-fold lower potency at -opioid receptor;37 120-fold lower potency at opioid receptor37 Intravenous, subcutaneous, and oral (including enteric-coated) 675 (SD 495855) (03 mg/kg*)38 Peripheral -opioid receptor antagonist; 30-fold lower potency at -opioid receptor;37 30-fold lower potency at -opioid receptor37 Oral 957 (12 mg);37,39 113 (SD 68158) (12 mg);37,40 507158 (6, 12, 18, and 24 mg)37,41 15 (12 mg)37,39; 21 (12 mg)37,40; 30 (12 mg, steady-state values)37,42 461 (SD 191) (12 mg)37,40 2455 (6, 12, 18, and 24 mg)37,41 Gut ora41,44 ADL-08-0011, an amide hydrolysis product that is generated by gut ora41,44

Route Cmax (ng/mL)

tmax (h) AUC0- (ngh1mL1) t1/2 (h) Metabolism Active metabolite Elimination

010 (SD 005115) (03 mg/kg*)38 353 (SD 262444) (03 mg/kg*)38 29 (SD 2038) (03 mg/kg*)38 Small percentage undergoes hepatic metabolism (possibly glucuronidation)43 None

Mostly eliminated by renal excretion, about Major excretion faecal, minor excretion urine16 4050% excreted unchanged in urine45

AUC=area under the curve. *Every 6 h as 12 intravenous doses. Healthy volunteers. Single oral dose. Patients older than 65 years. Patients with chronic obstipation.

Table 1: Comparison of pharmacological data for methylnaltrexone and alvimopan

1200

New Drug Class

Design Healthy volunteers Yuan et al (1996)51 Phase I/II, randomised, double-blind, placebo-controlled, crossover Phase I/II, randomised double-blind, placebo-controlled, crossover Phase II, non-randomised, single-blind, dose-escalating Phase II, randomised, double-blind, placebo-controlled

n 12

Intervention Group 1: placebo; group 2: placebo+005 mg/kg morphine; group 3: 005 mg/kg morphine+045 mg/kg methylnaltrexone; all intravenously Group 1: placebo; group 2: 009 mg/kg morphine; group 3: 009 mg/kg morphine+03 mg/kg methylnaltrexone; all intravenously 064 mg/kg, 64 mg/kg, and 192 mg/kg oral methylnaltrexone Group 1: oral placebo+intravenous placebo; group 2: oral placebo+005 mg/kg intravenous morphine; group 3: 192 mg/kg oral methylnaltrexone +005 mg/kg intravenous morphine

Result Oral-caecal transit time (min, mean [SD]): group 1: 1046 (311); group 2: 1633 (398) (p<001 vs group 1); group 3: 1063 (398) (not signicant vs group 1, p=056 vs group 2); methylnaltrexone did not antagonise analgesic eect of morphine Time for gastric volume to decrease by 50% (min, mean [SD]): group 1: 55 (21); group 2: 210 (90) (p<003 vs group 1); group 3: 74 (30) (p<004 vs group 2) Methylnaltrexone was safe and well tolerated up to maximum dose Oral-caecal transit (min, mean [SD]): group 1: 1146 (370); group 2: 1585 (502) (p<0001 vs group 1); group 3: 1104 (450) (not signicant vs group 1, p<0005 vs group 2)

Murphy et al (1997)52 Yuan et al (1997)53 Yuan et al (1997)53

11

14 14

Patients with chronic methadone-induced constipation Yuan et al (2000)54 Yuan et al (2000)55 Phase II, randomised, double-blind, placebo-controlled 22 Decrease from baseline in oral-caecal transit time (min, mean [SD]): group 1: Group 1: placebo; group 2: 00150365 mg/kg methylnaltrexone; all intravenously; both groups tested 14 (12); group 2: 777 (372) (p<001 vs group 1); no laxation with group 1 vs laxation for all of group 2 (p<001); no opioid withdrawal symptoms on days 1 and 2 Placebo followed next day with: group 1: 03 mg/kg methylnaltrexone; group 2: 10 mg/kg methylnaltrexone; group 3: 30 mg/kg methylnaltrexone; all orally Time to bowel movement (h, mean [SD]): group 1 (results for 3 of 4 patients): 180 (87); group 2 (4 of 4 patients): 123 (87); group 3 (4 of 4 patients): 52 (45); dose-response eect with drug, p=004; no adverse eects; no opioid withdrawal symptoms Proportion of patients that laxated within 4 h of rst dose: group 1: 484%; group 2: 155%; 33% dierence (95% CI 7%49%; p<00001); no opioid withdrawal symptoms

Phase II, single-blind, dose-ranging 12

Patients with advanced illness and opioid-induced bowel dysfunction Thomas et al (2008)56 Viscusi et al (2005)57 Phase III, randomised, double-blind, placebo-controlled 133 Group 1: placebo; group 2: 015 mg/kg methylnaltrexone; all subcutaneously every other day, for 2 weeks

Patients with postoperative ileus 65 Phase II, randomised, double-blind, placebo-controlled, 90 min after segmental colectomy Time to rst bowel movement (h, mean [SD]): group 1: 120 (10); group 2: 97 Group 1: placebo; group 2: 03 mg/kg methylnaltrexone; every 6 h by intravenous infusion for (6) (p=001 vs group 1); time to discharge eligibility (h, mean [SD]): group 1: 149 (17); group 2: 119 (7) (p=003 vs group 1) 7 days or up to 24 h after gastrointestinal recovery

Table 2: Clinical studies with methylnaltrexone

Additionally the drug has several interesting features when given intravenously,66 which need further eludication: reduction of opioid-induced urinary retention,67 reversal of opioid-induced cough reex suppression,68 and inhibition of both opioid-induced angiogenesis and opioid-receptor-independent vascular endothelial growth factor receptor transactivation.69

of methylnaltrexone in larger populations than those previously studied.

Alvimopan
Pharmacology
After peripherally-acting methylnaltrexone was shown to be eective, the synthetic molecule alvimopan was developed (gure 3). The drug (Entereg) was initially reported in 1994 by Eli Lilly71 and is presently being co-developed by Adolor Corporation, Exton, PA, USA, and GlaxoSmithKline. Alvimopan is a quarternary -opioid receptor antagonist with a dierent pharmacological prole from methylnaltrexone (table 1).72 The high-molecular-weight zwitterionic form and polarity restrict gastrointestinal absorption and prevent the drug crossing the bloodbrain barrier.7 Compared with methylnaltrexone, in-vitro data show that alvimopan has a higher binding anity for human -opioid receptors and is more potent.73,74 Unlike methylnaltrexone, alvimopan has an active metaboliteADL-08-0011which is produced by amide hydrolysis in human gut ora,41,44 and seems to be absorbed systemically. In vitro, the metabolite is equipotent to alvimopan, but its contribution to clinical eectiveness is unknown.44 The pharmacokinetic prole of the drug is characterised by an oral bioavailability of only 6%, which results in
1201

Adverse eects
Pharmacokinetic study of the safety prole of methylnaltrexone identied the dose-limiting adverse eectorthostatic hypotensionat plasma concentrations in excess of 1400 ng/mL; the eect was transient and self-limiting. Consequently, an upper dose limit of 03 mg/kg was chosen for further clinical assessment,53 a dose that is well tolerated in healthy volunteers, in members of methadone maintenance programmes, and in dierent clinical populations. The most common adverse eects are abdominal cramps and atulence.38,54,70 In a phase III placebo-controlled trial in 133 patients with opioid-induced bowel disorder and advanced illnesses, abdominal pain was reported by 17% of patients (vs 13% of placebo group), atulence by 13% of patients (vs 7% of placebo group), and nausea by 11% of patients (vs 7% of placebo group).56 Rare adverse events might only be seen with long-term use
www.thelancet.com Vol 373 April 4, 2009

New Drug Class

OH O HO H3C CH3 N H O

exploratory post-hoc analysis15 showed that in the subgroup receiving opioids for postoperative pain via intravenous patient-controlled analgesia, statistically gastrointestinal recovery was signicantly faster in the treated group than in the placebo group. In this European and New Zealand study,15 69% of patients were given non-opioid analgesics in the rst 48 h after surgery. Bchler and colleagues15 report that the proportion of patients in the US and Canadian studies of alvimopan in postoperative ileus,8183 in which most patients used patient-controlled analgesia, was substantially lower at only 4%. Therefore, the response of patients treated with alvimopan via patient-controlled analgesia in the European and New Zealand study15 is consistent with the ndings in the US and Canadian trials,8183 with respect to recovery of gastrointestinal function.

Figure 3: Chemical structure of anhydrous alvimopan Structure taken from Martindale.

Adverse eects
McNicol60 also analysed the safety data from nine placebo-controlled trials of alvimopan in healthy volunteers,7577 in patients given chronic opioid therapy,78 or in patients after surgery who were at risk of postoperative ileus.8084 Surprisingly, the proportion of gastrointestinal-related adverse events was lower in patients treated with alvimopan than with placebo, which suggests that gastrointestinal outcomes are measures of ecacy rather than safety.60 As might be expected, most studies suggested a reduction of the prevalence and severity of constipation, postoperative ileus, and vomitingeg, one study showed a substantial reduction of nausea in 26 patients post-surgery who were given 6 mg of alvimopan daily.80 However, statistically signicant reduction of nausea with alvimopan was not seen by meta-analysis with a xed-eect model as an overall class eect.60 Similar to data suggesting that intravenous methylnaltrexone reduces opioid-induced urinary retention,67 meta-analysis60 of two studies of alvimopan in patients after surgery81,83 showed reduced reports of oliguria with an absolute risk reduction of 4% (95% CI 7 to 1; ie, not formally statistically signicant). A study in patients on chronic opioid therapy with opioid-induced bowel dysfunction78 showed an increased prevalence of diarrhoea in patients treated with alvimopan, with an absolute risk increase of 11% (95% CI 2 to 19), but there was no dierence in the other populations studied or in those given methylnaltrexone. One study79 suggested that for patients with non-cancer pain and opioid-induced bowel dysfunction, 05 mg of alvimopan twice daily had the best benet-to-risk prole. Generally, more safety data are available for alvimopan than for methylnaltrexone,60 but both seem to be well tolerated. These promising results made it probable that the FDA would approve the 12 mg dose of alvimopan to treat postoperative ileus. However, the letter of approval, issued in November, 2006, called for additional safety data and a risk-management plan12 because of the results from a 12-month study.86 The study86 was designed to
www.thelancet.com Vol 373 April 4, 2009

predominant activity on the gut itself.16 Alvimopan does not seem to be metabolised by cytochrome P450, glucuronidation, or sulphation, and the major route of excretion is faecal.16 Pharmacokinetic study of patients with renal impairment who were given a single oral dose of 12 mg showed that there was no relation between renal function and plasma alvimopan.40

Ecacy
Preclinical studies showed that alvimopan is a potent antagonist of opioid-mediated inhibition of gastrointestinal function, without antagonising CNS-mediated eects such as analgesia or pupillary constriction.16,37 Eects in animal models have been conrmed in healthy volunteers,75,76 and in patients given opioid therapy for chronic pain or methadone addiction (table 3).70,72 The ecacy of alvimopan in opioid-induced bowel dysfunction has been shown in phase II/III clinical studies, in which median time to rst bowel movement decreased signicantly78 and mean weekly bowel movement increased signicantly.79 McNicols60 systematic review assessed the ecacy of alvimopan for the prevention of postoperative ileus from ve placebo-controlled trials, which enrolled a total of 2225 patients at risk.8084 The combined outcomes showed that time to rst bowel movement or atus, or total time to tolerate solid food and pass atus or stool, was quicker with alvimopan than with placebo. Increase of the daily dose from 6 mg to 12 mg did not oer an advantage, and no dierence in pain scores (visual analogue scale) was reported between alvimopan and placebo. 39 (4%) of 956 patients given alvimopan compared with 63 (10%) of 648 patients given placebo reported having ileus, translating to an absolute risk reduction of 4% (95% CI 7 to 1). A large multicentre trial in Europe and New Zealand15 showed a reduction of the mean time to tolerate solid food and the time to either rst atus or bowel movement in patients with postoperative ileus who were treated with alvimopan. By contrast with US and Canadian studies,8183 however, this eect was not statistically signicant.15 An
1202

New Drug Class

assess the long-term safety and tolerability of alvimopan in patients taking opioids for chronic non-cancer pain and who had opioid-induced bowel dysfunction. Of 805 patients randomised 2:1538 given alvimopan (05 mg twice daily) and 267 given placebothe proportion of serious adverse events was similar in both

groups (13% vs 11%), but there was a numerical imbalance in the proportion of neoplasms (28% vs 07%) and an increased proportion of myocardial infarctions (n=7 [13%] vs n=0). However, the serious cardiovascular adverse events arose in patients with established or high risk of cardiovascular disease and, since most events took

Design Healthy volunteers Liu et al (2001)75 Phase I/II, randomised, double-blind, placebo-controlled, crossover

Intervention

Result

14

Group 1: oral placebo+intravenous placebo; group 2: Gastrointestinal transit time (min, mean [SD]): group 1: 69 (33); group 2: 103 (37); group 3: 76 (30) (p=0004 vs group 2) oral placebo+005 mg/kg intravenous morphine; group 3: 4 mg oral alvimopan+005 mg/kg intravenous morphine; washout period of 5 days Group 1: 4 mg oral alvimopan+015 mg/kg intravenous morphine; group 2: oral placebo+015 mg/kg intravenous morphine; group 3: oral placebo+intravenous placebo Group 1: 30 mg morphine+placebo; group 2: 30 mg morphine+3 mg alvimopan; all orally for 4 days, morphine twice a day, placebo or alvimopan three times a day, washout period 10 days Group 1: placebo; group 2: 05 mg alvimopan; group 3: 15 mg alvimopan; group 4: 30 mg alvimopan; all orally and single dose Group 1: placebo; group 2: 05 mg alvimopan; group 3: 10 mg alvimopan; all orally, once a day Group 1: placebo; group 2: 05 mg alvimopan twice a day; group 3: 10 mg alvimopan once a day; group 4: 10 mg alvimopan twice a day; all orally Pupil size, area under plasma concentration time curve; (mean [SD]); group 1: 453 (168); group 2: 423 (175); group 3: 44 (84) (p<0001 vs group 2; p<0001 vs group 1); categorical pain and pain relief scores showed signicant analgesia with morphine that was unaected by alvimopan Colonic motility (marker score with radio-opaque markers, mean [SD]); group 1: 248 (88); group 2: 140 (88) (p<001 vs group 1)

Liu et al (2001)75

45 Phase II, randomised, double-blind, placebo-controlled study (healthy volunteers after dental surgery) Phase II, randomised, double-blind, placebo-controlled, crossover Phase II, randomised, double-blind, placebo-controlled Phase II/III, randomised, double-blind, placebo-controlled Phase II/III, randomised, double-blind, double-dummy, placebo-controlled Phase III, randomised, double-blind, placebo-controlled 11

Barr et al (2000)77

Patients with chronic methadone-induced constipation or opioid-induced bowel dysfunction Schmidt (2001)72 Paulson et al (2005)78 Webster et al (2008)79 75 Proportion of patients with bowel movement within 12 h: group 1: 30%; group 2: 65%; group 3: 77%; group 4: 100%; (p<0001 for overall treatment with alvimopan) Time to rst bowel movement (h, median): group 1: 21; group 2: 7; group 3: 3 (hazard ratio vs group 1 229 [95% CI 155339; p<0001]) Change in spontaneous bowel movement frequency compared with group 1 (weekly, mean [95% CI]): group 2: 171 (083258); group 3: 164 (088240); group 4: 252 (140365); (p<0001 for all comparisons)

168

522

Patients with postoperative ileus Bchler et al (2008)15 911 Group 1: oral placebo; group 2: 6 mg alvimopan; group 3: 12 mg alvimopan; all orally twice a day; opioids for postoperative pain were administered as patient-controlled analgesia or bolus injection Time to recovery of gastrointestinal functions (composite endpoint) (h, mean [SE]): group 1: 926 (306); group 2: 842 (237); group 3: 878 (268); dierence in time to recovery of gastrointestinal function compared with group 1 (h, mean [95% CI]): group 2: 85 (160 to 09) (p=0042) (non-signicant because threshold was p<0025 for application of Hochbergs method to correct for multiple comparisons); group 3: 48 (128 to 32) (p=020); an exploratory post-hoc analysis showed that alvimopan was more eective in the patient-controlled analgesia group than in the bolus injection group Hazard ratio of time to rst bowel movement compared with group 1:(95% CI); group 2: 12 (0526); group 3: 29 (1366); (p=001 for both comparisons) Hazard ratio of time to recovery of gastrointestinal functions (composite endpoint) compared with group 1 (95% CI): group 2: 128 (100164) (p<005); group 3: 154 (121196) (p<0001) Hazard ratio of time to recovery of gastrointestinal functions (composite endpoint) compared with group 1 (95% CI): group 2: 145 (113185) (p=0003); group 3: 128 (099164) (p=0059) Hazard ratio of time to recovery of gastrointestinal functions (composite endpoint) compared with group 1 (95% CI): group 2: 124 (101153) (p=0037); group 3: 126 (103154) (p=0028); results adjusted for covariateseg, sex or surgical duration Time to rst bowel movement and toleration of solid food (composite endpoint) (h, mean): group 1: 920; group 2: 718; dierence 202 h (95%CI 265 to 139); alvimopan accelerated time to rst bowel movement (hazard ratio 233, p<0001); similar proportion of adverse events: 65% in group 1, 56% in group 2 Hazard ratio of time to rst bowel movement and toleration of solid food (composite endpoint) compared with group 1 (95% CI): 15 (129182) (p<0001)

Taguchi et al (2001)80 Wol et al (2004)81 Delaney et al (2004)82 Viscusi et al (2001)83 Herzog et al (2006)84 Kirk et al (2008)85

Phase III, randomised, double-blind, placebo-controlled Phase III, randomised, double-blind, placebo-controlled Phase III, randomised, double-blind, placebo-controlled Phase III, randomised, double-blind, placebo-controlled Phase III, randomised, double-blind, placebo-controlled Phase III, randomised, double-blind, placebo-controlled

79

Group 1: placebo; group 2: 1 mg alvimopan twice a day; group 3: 6 mg alvimopan twice a day; all orally Group 1: placebo; group 2: 6 mg alvimopan; group 3: 12 mg alvimopan; all orally twice a day Group 1: placebo; group 2: 6 mg alvimopan; group 3: 12 mg alvimopan; all orally twice a day Group 1: placebo; group 2: 6 mg alvimopan; group 3: 12 mg alvimopan; all orally twice a day Group 1: placebo; group 2: 12 mg alvimopan; all orally twice a day

510

451

615

519

654

Group 1: placebo; group 2: 12 mg alvimopan orally twice a day; both groups followed a standardised accelerated postoperative care pathway (early ambulation, oral feeding, and nasogastric tube removal) to help with gastrointestinal tract recovery

Table 3: Clinical studies with alvimopan

www.thelancet.com Vol 373 April 4, 2009

1203

New Drug Class

place within the rst 12 weeks of treatment, these events did not seem to be linked to the duration of dosing.86 Adolor Corporation submitted a revised risk-management programme for alvimopan in February, 2008, and alvimopan was approved in May, 2008, for the management of postoperative ileus with a Risk Evaluation and Mitigation Strategy (REMS).87 REMS is designed to ensure that the benets of the drug outweigh the risks and includes: restriction to inpatient use only, hospitals should be specially certied, educational material should be distributed to health-care professionals, and the eectiveness of the REMS should be regularly assessed.

Conclusions
Methylnaltrexone and alvimopan are better than placebo for reversal of opioid-mediated increase of gastrointensinal transit time and constipation.60 At therapeutic intravenous doses of 03045 mg/kg and oral doses up to 19 mg/kg, methylnaltrexone is well tolerated and able to relieve constipation in methadone-dependent individuals and patients with advanced illnesses who need high doses of opioids for pain control.48,88 Consequently, the FDA and the European Medicines Agency have approved subcutaneous methylnaltrexone for treatment of opioid-induced constipation in adults with advanced illnesses. Methylnaltrexone should be used in patients with opioid-induced bowel dysfunction who do not have a response to a reasonable laxative regimen, in combination with the laxative regimen. The recommended dose is 8 mg for patients weighing 3861 kg and 12 mg for patients weighing 62114 kg, every 2 days. Outside these weight ranges, the dose should be 015 mg/kg.74 Defaecation can be expected within 4 h after the rst dose in about 50% of patients.56 Alvimopan is eective in patients with postoperative ileus at doses of 6 mg or 12 mg daily and the drug is approved by the FDA for accelerated restoration of normal bowel function in patients aged 18 years and over who have undergone partial resection surgery on the large or small bowel. The recommended dose is one 12 mg capsule just before surgery and another 12 mg dose twice daily for up to 7 days, but not exceeding 15 doses. Despite the high probability of eectiveness, the use of both drugs will be limited by expense. From personal communication with the manufacturers, we are informed that a single dose of 12 mg methylnaltrexone costs about 5055 in Europe and the market price for one single dose of 12 mg alvimopan in the USA is US$5250 (US$1=075). Cost-eectiveness studies investigating the incremental costs per quality-adjusted life-year are needed, as suggested by the UK National Institute for Health and Clinical Excellence.89 Presently, direct comparison between methylnaltrexone and alvimopan is not possible, because such largescale clinical trials have not yet been done. Alvimopan seems to have higher pharmacological potency than
1204

methylnaltrexone,73,74 but methylnaltrexone can be given via dierent routes, which might be benecial for early postoperative or terminally ill patients, whereas alvimopan is available only orally. Methylnaltrexone also seems to have positive intrinsic -opioid receptor functional activity, consistent with partial antagonism, unlike alvimopan and ADL-08-0011, which have negative intrinsic activities.90 Conceivably, an inverse agonist would produce a larger clinical eect than a partial agonist because of a more complete reversal of action of highly ecacious agonists such as morphine. But the validity of this hypothesis and the clinical signicance of these dierences remain to be determined. Other potentially relevant clinical dierences between alvimopan and methylnaltrexone should be investigated in future clinical trials. Most studies of methylnaltrexone and alvimopan measure ecacy. The treated and control groups are homogeneous, confounding variables are controlled, and bias is reduced. Therefore, internal validity of the studies is comparatively high, although some studies used a pseudo-intention-to-treat principle rather than a real intention-to-treat analysis. But the external validity of the studies to the general population of patientseg, those with canceris low. To assess eectiveness, there is a need for studies assessing clinical practice and reecting real-life circumstances. Larger trials are needed to examine the eects of opioid antagonists on pain control, to determine proper dosing, and to compare opioid antagonist ecacy with standard adjuvant therapy. Furthermore, early postoperative patients or those with advanced illnesses often receive several drugs and the possible pharmacological interactions or the eects of enzyme induction by comedications are unknown and need future investigation in large groups of patients and clinical trials.
Conict of interest statement We declare that we have no conict of interest. Acknowledgments We thank Sabine Buroh, librarian at the University Hospital Freiburg, for her assistance with the electronic literature search, and Erika Graf, statistician from the clinical trials centre in the University Hospital Freiburg, for her assistance with calculating the 95% CI in some studies. References 1 Panchal SJ, Mller-Schwefe P, Wurzelmann JI. Opioid-induced bowel dysfunction: prevalence, pathophysiology and burden. Int J Clin Pract 2007; 61: 118187. 2 Pappagallo M. Incidence, prevalence, and management of opioid bowel dysfunction. Am J Surg 2001; 182 (suppl 5a): 11S18S. 3 Sykes NP. The relationship between opioid use and laxative use in terminally ill cancer patients. Palliat Med 1998; 12: 37582. 4 Schug SA, Zech D, Grond S, Jung H, Meuser T, Stobbe B. A long-term survey of morphine in cancer patients. J Pain Symptom Manage 1992; 7: 25966. 5 Kurz A, Sesser I. Opiate-induced bowel dysfunction: pathophysiology and potential new therapies. Drugs 2003; 63: 64971. 6 McNicol E, Horowicz-Mehler N, Fisk RA, et al. Management of opioid side-eects in cancer-related and chronic noncancer pain: a systematic review. J Pain 2003; 4: 23156. 7 Thomas J. Opioid-induced bowel dysfunction. J Pain Symptom Manage 2008; 35: 10313.

www.thelancet.com Vol 373 April 4, 2009

New Drug Class

9 10 11 12 13 14

15

16 17

18 19 20

21

22

23 24 25

26

27 28

29 30

31

32

33 34

Fakata KL, Cole BE. Peripheral opioid antagonists: a therapeutic advance for optimizing opioid gastrointestina tolerability. J Fam Pract 2007; 56 (suppl 6): S312. Stewart D, Waxman K. Management of posteroperative ileus. Am J Ther 2007; 14: 56166. Kehlet H, Holte K. Review of postoperative ileus. Am J Surg 2001; 182 (suppl 5A): 3S10S. Leslie JB. Alvimopan: a peripherally acting mu-opioid receptor antagonist. Drugs Today (Barc) 2007; 43: 61125. Kraft MD. Emerging pharmacologic options for treating postoperative ileus. Am J Health Syst Pharm 2007; 64: S1320. Bauer AJ, Boeckxstaens GE. Mechanisms of postoperative ileus. Neurogastroenterol Motil 2004; 16 (suppl 2): 5460. Delaney CP, Wol BG, Viscusi ER, et al. Alvimopan for postoperative ileus following bowel resection. Ann Surg 2007; 245: 35563. Bchler MW, Sieler CM, Monson JRT, et al. Clinical trial: alvimopan for the management of post-operative ileus after abdominal surgery: results of an international randomized, double-blind, multicentre, placebo-controlled clinical study. Aliment Pharmacol Ther 2008; 28: 31225. Camilleri M. Alvimopan, a selective peripherally acting -opioid antagonist. Neurogastroenterol Motil 2005; 17: 15765. Gomez-Flores R, Rice KC, Zhang X, Weber RJ. Increased tumor necrosis factor- and nitric oxide production by rat macrophages following in vitro stimulation and intravenous administration of the -opioid agonist SNC 80. Life Sci 2001; 68: 267584. Kowalski J. Augmenting eect of opioids on nitrite production by stimulated murine macrophages. Neuropeptides 1998; 32: 28791. Gomez-Flores R, Weber RJ. Immunomodulation of macrophage functions by opioids. Adv Exp Med Biol 1998; 437: 1319. Narayan P, Singh VK, Agarwal SS, et al. Immunmodulation by opioid peptidomimetic compound. Neuroimmunomodulation 2001; 9: 13440. Kal JC, Schraut WH, Billiar TR, Simmons RL, Bauer AJ. Role of inducible nitric oxide synthase in postoperative intestinal smooth muscle dysfunction in rodents. Gastroenterology 2000; 118: 31627. Traut U, Brgger L, Pauli-Magnus C, et al. Systemic prokinetic pharmacologic treatment for postoperative adynamic ileus following abdominal surgery in adults. Cochrane Database Syst Rev 2008; 1: CD004930. Behm B, Stollmann N. Postoperative ileus: etiologies and interventions. Clin Gastroenterol Hepatol 2003; 1: 7180. Peeters TL. New motilin antagonists: a long and winding road. Neurogastroenterol Motil 2006; 18: 15. Tuteja AK, Rao SS. Lubiprostone for constipation and irritable bowel syndrome with constipation. Expert Rev Gastroenterol Hepatol 2008; 2: 72733. Saad R, Chey WD. Lubiprostone for chronic idiopathic constipation and irritable bowel syndrome with constipation. Expert Rev Gastroenterol Hepatol 2008; 2: 497508. Maron DJ, Fry RD. New therapies in the treatment of postoperative ileus after gastrointestinal surgery. Am J Ther 2008; 15: 5965. Saunders WB, Bowers B, Moss B, Bell TJ, Wang PF. Recorded rate and economic burden associated with postoperative ileus. 39th American Society of Health-System Pharmacists Midyear Clinical Meeting and Exhibits; Orlando, FL; Dec 59, 2004. Abstr P440E. Wittbrodt E. The impact of postoperative ileus and emerging therapies. Pharm Treat 2006; 31: 3959. Becker G, Galandi D, Blum HE. Peripherally acting opioid antagonists in the treatment of opiate-related constipation: a systematic review. J Pain Symptom Manage 2007; 3: 54765. Friedman JD, Bello Buono FA. Opioid antagonists in the treatment of opiate-induced constipation and pruritus. Ann Pharmacother 2001; 35: 8591. Knowles CH, Martin JE. Slow transit constipation: a model of human gut dysmotility. Review of possible aetiologies. Neurogastroenterol Motil 2000; 12: 18196. Holzer P. Treatment of opioid-induced gut dysfunction. Expert Opin Investig Drugs 2007; 16: 18194. Sternini C, Patierno S, Selmer IS, Kirchgessner A. The opioid system in the gastrointestinal tract. Neurogastroenterol Motil 2004; 16 (suppl 2): 316.

35

36

37

38

39

40

41

42

43

44

45 46 47

48

49

50

51

52

53

54

55 56

57

De Schepper HU, Cremonini F, Park MI, Camilleri M. Opioids and the gut: pharmacology and current clinical experience. Neurogastroenterol Motil 2004; 16: 38394. Neyens R, Jackson KC. Novel opioid antagonists for opioid-induced bowel dysfunction and postoperative ileus. J Pain Palliat Care Pharmacother 2007; 21: 2733. DeHaven-Hudkins DL, DeHaven RN, Little PJ, Techner LM. The involvement of the -opioid receptor in gastrointestinal pathophysiology: therapeutic opportunities for antagonism at this receptor. Pharmacol Ther 2008; 17: 16287. Yuan CS, Doshan H, Charney MR, et al. Tolerability, gut eects, and pharmacokinetics of methylnaltrexone following repeated intravenous administration in humans. J Clin Pharmacol 2005; 45: 53846. Foss J, Marbury TC, Melikian A, Schmith V, Du W, Wallin B. Pharmacokinetics and safety of alvimopan, a novel, oral peripherally acting mu-opioid receptor (PAM-OR) antagonist in the elderly. Pharmacotherapy 2005; 25: 1505 (abstr). Foss J, Marbury TC, Melikian A, Schmith V, Du W, Wallin B. Pharmacokinetics and safety of alvimopan, a novel, oral peripherally acting mu-opioid receptor (PAM-OR) antagonist in patients with renal impairment. Pharmacotherapy 2005; 25: 1504 (abstr). Foss J, Schmith V, Wallin B, Du W, Melikian A. Alvimopan (Entereg), a novel opioid antagonist, achieves active systemic concentrations. Clin Pharmacol Ther 2005; 77: P74 (abstr). Schmith V, Garnett W, Barr WH, et al. Alvimopan pharmacokinetics (PK) and pharmacodynamics (PD) in patients with chronic constipation. Clin Pharmacol Ther 2005; 77: P49 (abstr). Yuan CS, Israel RJ. Methylnaltrexone, a novel peripheral opioid receptor antagonist for the treatment of opioid side-eects. Expert Opin Investig Drugs 2006; 15: 54152. Foss JF, Fisher DM, Schmith VD. Pharmacokinetics of alvimopan and its metabolite in healthy volunteers and patients in postoperative ileus trials. Clin Pharmacol Ther 2008; 83: 77076. De Ponti F. Methylnaltrexone. Curr Opin Investig Drugs 2002; 3: 61420. Brown DR, Goldberg LI. The use of quaternary narcotic antagonists in opiate research. Neuropharmacology 1985; 24: 18191. Russell J, Bass P, Goldberg LI, Schuster CR, Merz H. Antagonism of gut, but not central eects of morphine with quaternary narcotic antagonists. Eur J Pharmacol 1982; 78: 25561. Yuan CS, Israel RJ. Methylnaltrexone, a novel peripheral opioid receptor antagonist for the treatment of opioid side-eects. Expert Opin Investig Drugs 2006; 15: 54152. Kotake AN, Kuwahara SK, Burton E, McCoy CE, Goldberg LI. Variations in demethylation of N-methylnaltrexone in mice, rats, dogs, and humans. Xenobiotica 1989; 19: 124754. Holzer P. Opioids and opioid receptors in the enteric nervous system: from a problem in opioid analgesia to a possible new prokinetic therapy in humans. Neurosci Lett 2004; 361: 19295. Yuan CS, Foss JF, OConnor M, Toledano A, Roizen MF, Moss J. Methylnaltrexone prevents morphine-induced delay in oral-cecal transit time without aecting analgesia: a double-blind randomized placebo-controlled trial. Clin Pharmacol Ther 1996; 59: 46975. Murphy DB, Sutton JA, Prescott LF, Murphy MB. Opiate-induced delay in gastric emptying: a peripheral mechanism in humans. Anesthesiology 1997; 87: 76570. Yuan CS, Foss JF, Osinski J, Toledano A, Roizen MF, Moss J. The safety and ecacy of oral methylnaltrexone in preventing morphine-induced delay in oral-cecal transit time. Clin Pharmacol Ther 1997; 61: 46775. Yuan CS, Foss JF, OConnor M, et al. Methylnaltrexone for reversal of constipation due to chronic methadone use. JAMA 2000; 283: 36772. Yuan CS, Foss JF. Oral methylnaltrexone for opiate-induced constipation. JAMA 2000; 284: 138384. Thomas J, Karver S, Cooney GA, et al. Methylnaltrexone for opioid-induced constipation in advanced illness. N Engl J Med 2008; 358: 233243. Viscusi E, Rathmell J, Fichera A, Gan TL, Israel RJ. A double-blind, randomized, placebo-controlled trial of methylnaltrexone (MNTX) for post-operative bowel dysfunction in segmental colectomy. Proc Am Soc Anesth 2005; 103: A893 (abstr).

www.thelancet.com Vol 373 April 4, 2009

1205

New Drug Class

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72 73

74

Yuan CS, Foss JF, OConnor M, et al. Eects of enteric-coated methylnaltrexone in preventing opioid-induced delay in oral-cecal transit time. Clin Pharmacol Ther 2000; 67: 398404. Yuan CS, Wei G, Foss JF, OConnor M, Karrison T, Osinski J. Eects of subcutaneous methylnaltrexone on morphine-induced peripherally mediated side-eects: a double-blind randomized placebo-controlled trial. J Pharmacol Exp Ther 2002; 300: 11823. McNicol ED, Boyce D, Schumann R, Carr DB. Mu-opioid antagonists for opioid-induced bowel dysfunction. Cochrane Database Syst Rev 2008; 2: CD006332. Thomas J, Portenoy R, Moehl M, et al. A phase II randomized dose-nding trial of methylnaltrexone for the relief of opiate-induced constipation in hospice patients. Proc Am Soc Clin Oncol 2003; 22: 729 (abstr 2933). Portenoy R, Thomas J, Moehl Boatwright ML, et al. Subcutaneous methylnaltrexone for the treatment of opioid-induced constipation in patients with advanced illness: a double-blind, randomized, parallel group, dose-ranging study. J Pain Symptom Manage 2008; 35: 45868. Thomas J, Lipman A, Slatkin N, et al. A phase III double-blind placebo-controlled trial of methylnaltrexone (MNTX) for opioid-induced constipation (OIC) in advanced medical illness (AMI). ASCO Annual Meeting; Orlando, FL; May 1317, 2005. Abstr 8003. http://opd.asco.org/ASCO/Abstracts+%26+Virtual+Meeting/ Abstracts?&vmview=abst_detail_view&confID=34&abstractID=32585 (accessed Jan 10, 2009). Karver S, Israel RJ. Methylnaltrexone improves opioid-induced constipation and reduces laxative use in patients with advanced illness. Gastroenterology 2007; 132 (suppl 2): A478 (abstr). Wyeth Pharmaceuticals. Wyeth and Progenics begin phase II and III trials of methylnaltrexone for OIC and POI. Madison, NJ: Wyeth Pharmaceuticals. Reichle FM, Conzen PF. Methylnaltrexone, a new peripheral -receptor antagonist for the prevention and treatment of opioid-induced extracerebral side-eects. Curr Opin Investig Drugs 2008; 9: 90100. Rosow CE, Gomery P, Chen TY, Stefanovich P, Stambler N, Israel R. Reversal of opioid-induced bladder dysfunction by intravenous naloxone and methylnaltrexone. Clin Pharmacol Ther 2007; 82: 4853. Foss JF, Orelind E, Goldberg LI. Eects of methylnaltrexone on morphine-induced cough suppression in guinea pigs. Life Sci 1996; 59: PL23538. Singleton PA, Lingen MW, Fekete MJ, Garcia JG, Moss J. Methylnaltrexone inhibits opiate and VEGF-induced angiogenesis: role of receptor transactivation. Microvasc Res 2006; 72: 111. Yuan CS, Foss JF, OConnor M, Osinski J, Roizen MF, Moss J. Eects of intravenous methylnaltrexone on opioid-induced gut motility and transit time changes in subjects receiving chronic methadone therapy: a pilot study. J Pain 1999; 83: 63135. Zimmerman DM, Gidda JS, Cantrell BE, Schoepp DD, Johnson BG, Leander JD. Discovery of a potent, peripherally selective trans-3,4-dimethyl-4(3-hydroxyphenyl) piperidine opioid antagonist for the treatment of gastrointestinal disorders. J Med Chem 1994; 37: 226265. Schmidt WK. Alvimopan (ADL 8-2698) is a novel peripheral opioid antagonist. Am J Surg 2001; 182: 27S38S. Goodman AJ, Le Bourdonnec B, Dolle RE. Mu opioid receptor antagonists: recent developments. ChemMedChem 2007; 2: 155270. Neary P, Delaney CP. Alvimopan. Expert Opin Investig Drugs 2005; 14: 47988.

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

Liu SS, Hodgson PS, Carpenter RL, Fricke JR Jr. ADL 8-2698, a trans-3,4-dimethyl-4-(3-hydroxyphenyl) piperidine, prevents gastrointestinal eects of intravenous morphine without aecting analgesia. Clin Pharmacol Ther 2001; 69: 6671. Gonenne J, Camilleri M, Ferber I, et al. Eect of alvimopan and codeine on gastrointestinal transit: a randomized controlled study. Clin Gastroenterol Hepatol 2005; 3: 78491. Barr WH, Nguyen P, Slattery M, Russell R, Carpenter RL. ADL8-2698 reverses opioid induced delay in colonic transit. Clin Pharmacol Ther 2000; 67: 91 (abstr). Paulson DM, Kenndy DT, Donovick RA, et al. Alvimopan: an oral, peripherally acting, -opioid receptor antagonist for the treatment of opioid-induced bowel dysfunctiona 21-day treatment randomized clinical trial. J Pain 2005; 6: 18492. Webster L, Jansen JP, Peppin J, et al. Alvimopan, a peripherally acting mu-opioid receptor (PAM-OR) antagonist for the treatment of opioid-induced bowel dysfunction: results from a randomised, double-blind, placebo-controlled, dose-nding study in subjects taking opioids for chronic non-cancer pain. Pain 2008; 137: 42840. Taguchi A, Sharma N, Saleem RM, et al. Selective postoperative inhibition of gastrointestinal opioid receptors. N Engl J Med 2001; 345: 93540. Wol BG, Michelassi F, Gerkin TM, et al, and Alvimopan Postoperative Ileus Study Group. Alvimopan, a novel, peripherally acting opioid antagonist results of a multicenter, randomized, double-blind, placebo-controlled, phase III trial of major abdominal surgery and posterative ileus. Ann Surg 2004; 240: 72835. Delaney CP, Weese JL, Hyman H, et al. Phase III trial of alvimopan a novel, peripherally acting, mu-opioid antagonist, for posteropertive ileus after major abdominal surgery. Dis Colon Rectum 2005; 48: 111429. Viscusi ER, Goldstein S, Witkowski T, et al. Alvimopan, a peripherally acting mu-opioid antagonist, compared with placebo in postoperative ileus after major abdominal surgery. Surg Endosc 2006; 20: 6470. Herzog TJ, Coleman RL, Guerrieri JP, et al. A double-blind, randomized, placebo-controlled phase III study of the safety of alvimopan in patients who undergo simple total abdominal hysterectomy. Am J Obstet Gynecol 2006; 195: 44553. Kirk L, Enker WE, Delaney CP, et al. Gastrointestinal tract recovery in patients undergoing bowel resection. Arch Surg 2008; 143: 1098105. P & T Community. Alvimopan studies halted after cardiovascular problems reported. http://www.formkit.com/Daily/DailyDetail. cfm?chosen=65062 (accessed Jan 9, 2009). US Food and Drug Administration. FDA approves Entereg to help restore bowel function during surgery. http://www.fda.gov/bbs/ topics/NEWS/2008/NEW01838.html (accessed Jan 11, 2009). Herbert MK, Holzer P. Standardized concept for the treatment of gastrointestinal dysmotility in critically ill patientscurrent status and future options. Clin Nutr 2008; 27: 2541. National Institute for Health and Clinical Excellence. Health Technology Appraisal: methylnaltrexone for the treatment of opioid-induced bowel dysfunction in advanced illness or palliative care. http://www.nice.org.uk/media/851/C6/18BowelDysfunction DraftScope.pdf (accessed Jan 11, 2009). Beattie DT, Cheruvu M, Mai N, et al. The in vitro pharmacology of the peripherally restricted opioid receptor antagonists, alvimopan, ADL 08-0011 and methylnaltrexone. Naunyn Schmiedebergs Arch Pharmacol 2007; 375: 20520.

1206

www.thelancet.com Vol 373 April 4, 2009

Series

Health in the Occupied Palestinian Territory 5 The health-care system: an assessment and reform agenda
Awad Mataria, Rana Khatib, Cam Donaldson, Thomas Bossert, David J Hunter, Fahed Alsayed, Jean-Paul Moatti

Attempts to establish a health plan for the occupied Palestinian territory were made before the 1993 Oslo Accords. However, the rst ocial national health plan was published in 1994 and aimed to regulate the health sector and integrate the activities of the four main health-care providers: the Palestinian Ministry of Health, Palestinian non-governmental organisations, the UN Relief and Works Agency, and a cautiously developing private sector. However, a decade and a half later, attempts to create an eective, ecient, and equitable system remain unsuccessful. This failure results from arrangements for health care established by the Israeli military government between 1967 and 1994, the nature of the Palestinian National Authority, which has little authority in practice and has been burdened by ineciency, cronyism, corruption, and the inappropriate priorities repeatedly set to satisfy the preferences of foreign aid donors. Although similar problems exist elsewhere, in the occupied Palestinian territory they are exacerbated and perpetuated under conditions of military occupation. Developmental approaches integrated with responses to emergencies should be advanced to create a more eective, ecient, and equitable health system, but this process would be dicult under military occupation.

Lancet 2009; 373: 120717 Published Online March 5, 2009 DOI:10.1016/S01406736(09)60111-2 This is the fth in a Series of ve papers on health in occupied Palestinian territory Institute of Community and Public Health, Birzeit University, Birzeit, occupied Palestinian territory (A Mataria PhD, R Khatib PhD); Department of Economics, Faculty of Commerce and Economics, Birzeit University, Birzeit, occupied Palestinian territory (A Mataria); Institute of Health and Society, Newcastle University, Newcastle-uponTyne, UK (Prof C Donaldson PhD); Department of Global Health and Population, Harvard School of Public Health, Cambridge, MA, USA (Prof T Bossert PhD); Centre for Public Policy and Health, School of Medicine and Health, Durham University, Teesside, UK (Prof D J Hunter PhD); Ministry of Health, Palestinian National Authority (F Alsayed MD); Institute of Health and Medical Research (INSERM) and Southeastern Health Regional Observatory (ORS PACA), Research Unit 912 and Faculty of Economics, University of the Mediterranean, Marseille, France (Prof J-P Moatti PhD) Correspondence to: Dr Awad Mataria, Institute of Community and Public Health, Birzeit University, Birzeit, PO Box 14, occupied Palestinian territory awad@birzeit.edu

Introduction
Achievement of the human right to the highest attainable standard of physical and mental health1 entails equitable access to eective health-care systems2 and calls for health professionals to promote necessary change.3 To strengthen the health systems of low-income and middle-income countries, the WHO Commission for Macroeconomics and Health calls for greater resources for health care.4 Since the signing of the Declaration of Principles on Interim Self-Government Arrangements by the Palestinian Liberation Organisation and the State of Israel, also known as the Oslo Accords,5 substantial donor assistance was meant to improve a health-care system for the occupied Palestinian territory (the West Bank, including the Palestinian Arab East Jerusalem, and the Gaza Strip). In 2003 alone, donations amounted to US$240 million ($65 per person), covering 87% of budgeted non-salary operating costs of the Ministry of Health.6 Health-care systems have three main goals: improving health, responding to the non-medical expectations of the population, and enhancing nancial risk protection.7,8 The rst and second reports in this Series9,10 on health and health services in the occupied Palestinian territory trace the steady improvement in the health status of the population until the mid-1990s when improvements slowed, and, in some cases, reversed. Reports two and three10,11 show how planning and coordination of health care are inadequate, the use of resources is ineective, and services are below acceptable standards, leading to public dissatisfaction with health services.12 Current nancial arrangements are associated with high risks and unequal burdens,13,14 with substantial out-of-pocket payments that favour rich people and place a high burden on poor people.15 Other reports in this Series911,16 elucidate the role of Israeli military occupation in producing and maintaining ineciencies and inequities and the relative
www.thelancet.com Vol 373 April 4, 2009

powerlessness of the Palestinian National Authority to counteract them. In this report, we use reviews of published work and interviews to identify ways to integrate developmental approaches with responses to emergencies to create a more eective, ecient, and equitable health system. We provide a prole of the Palestinian health-care system and analyse the system with respect to the six building blocks of the WHO framework8 for health systems: service delivery; workforce; information; medical products and technologies; nancing; and leadership, governance, and stewardship. We emphasise the complexity of health-system building under conditions of military occupation, review future political scenarios, and suggests ways to improve performance and equity.

Palestinian health-care system: a data prole

The 376 million Palestinians17 living in the occupied Palestinian territory are in the middle of epidemiological and demographic transitions.9,11 Four main providers18 are responsible for primary, secondary, and tertiary health care: a Palestinian Ministry of Health, Palestinian non-governmental organisations, the UN Relief and Works Agency,19 and the private sector. Health services are nanced through a mixture of taxes, health insurance premiums and co-payments, out-of-pocket payments, local community nancial and in-kind donations, and loans and grants from the international community (including the UN Relief and Works Agency).18 Reviews of the health sector estimated that total health expenditure in 2002 was 86% of gross domestic product (GDP)20 and per-person expenditure was $135 in 2005.6 Health, nutrition, and population indicators suggest that the occupied Palestinian territory is doing well compared with many countries in the region.9 Moreover, more than 95% of women receive some form of antenatal care and deliver in health institutions, and immunisation

1207

Series

coverage is high (>95%).21 However, socioeconomic and regional inequalities persist: an infant born to a family in the richest 20% in the West Bank is almost twice as likely to survive 1 year than is one born to a poor family in the Gaza Strip.14 Non-communicable diseases are the main causes of mortality;11 and, since 2000, there has been a substantial increase in the number of patients seeking mental-health services.22 In view of the turbulent situation, the chances of the occupied Palestinian territory achieving most Millennium Development Goals by 2015 are low.23 Until the Oslo Accords and the events of September, 2000,9 the occupied Palestinian territory had a functional, if dependent, economy. Lately, unemployment rates have spiralled upwards, and, in 2007, more than 215% of the active population was unemployed,24 leaving 572% of Palestinian households with an income less than the national poverty line of $318 per person per day.25 By 2006, GDP in the occupied Palestinian territory had fallen by a third of its 1999 value, from $1612 to $1129.26

Assessment of the health-care system

Complementarity between the four main providers of health care in occupied Palestinian territory has not developed from an attempt to establish a rational and ecient division of labour but has mainly arisen because of the political and economic situation. Closures,9 segregation,27 strikes,28 and impoverishment lead many transfers of patients from one provider to another.29 Restrictions on movement imposed by multiple checkpoints, barriers to movement,29,30 and the separation wall31,32 prevent access for patients and medical sta (gure). In July, 2007, alone, there were 40 recorded cases of ambulances being denied access to patients in the West Bank.34 A survey at the end of 2003 found that the number of people needing 1 h or more to reach an appropriate health facility had increased by ten times in 3 years.29 A network of mobile clinics now caters for the needs of people living in remote and isolated localities.35 There are few eective restrictions in the system, and patients seeking referral services are generally entitled to receive them.36 After the Oslo Accords, the number of primary healthcare centres run by non-governmental organisations fell from 242 in 1992 to 177 in 1994.37 The decrease was mainly due to abrupt changes in donor aid policies and the Palestinian National Authority budget allocation strategy,32,37,38 which aimed to reserve resources to rehabilitate the dilapidated services of the Ministry of Health. The increase in primary health-care centres sponsored by the Palestinian National Authority more than compensated for these losses, with about 170 new primary health-care facilities being opened (mostly in the West Bank) in 13 years.18 Moreover, in 2006, over 40 clinics were mutually operated by the Palestinian Ministry of Health and non-governmental organisations.39 Secondary and tertiary care are provided through general and specialised hospitals, mainly located in
1208

urban areas. The non-governmental sector operates 1582 beds in 28 hospitals (table), with a substantial proportion of its workload covered by the Palestinian National Authority health-insurance scheme.18 The shortage in tertiary health-care services is clear, and those concentrated in Jerusalem are inaccessible to most people in the occupied Palestinian territory because of Israeli restrictions on movement.30,31 As a result, an increasing number of cases are referred for treatment in other countriesabout 15 000 cases in 200521 mainly to Jordan, Egypt, and Israelthereby increasing the nancial burden on the system and society.41 The Ministry of Healths share in overall service delivery has risen signicantly, mainly because of severe population impoverishment, starting at the end of 2000, and the accompanying extension of free insurance coverage (see below).18 However, this increase was not matched by increased capacity of the ministry, resulting in falling quality of care. Most ministry hospitals have to turn away many patients because they have no capacity. By contrast, non-governmental hospitals are underused,18 emphasising the need for better coordination between providers. Despite several projects to promote quality of ministry-run services,20 they are still perceived as inferior.12 Financial accessibility to health services, especially for the most deprived sections of the population, has been compromised since 2000.28,29 Results of recent surveys show that a third of a representative sample of the population could not access health services because of high costs29 and that people living with nancial hardship or in poverty are twice as likely as rich people to be unsuccessful in accessing hospital care.42 Hence, patients preferences to improve quality of care have concentrated on urgently meeting the most basic needs.43 This pattern accords with Amartya Sens hypothesis of adaptive preferences:44 populations confronted with fundamental material constraints have objective diculty in expressing what their true needs would be if they had more opportunities available to them.45 Inequitable distribution of health facilities between and within the West Bank and the Gaza Strip, favouring the Gaza Strip and the central areas of the two regions, contributes to health inequityespecially under the exceptional restrictions on movement.30,32 The Gaza Strip has 14 beds for every 1000 people, and the West Bank has 12;18 in the West Bank, Ramallah has 11, whereas Salfeet district has only 02; and in the Gaza Strip, Gaza city has 21 and Rafah city only 05.18 Primary healthcare facilities, however, are more available in the West Bank than in the Gaza Strip18 (21 vs 09 centres per 10 000 people). This pattern is caused by more dispersed and enclosed populations on the West Bankdue to Israeli checkpoints and the separation walland results in ineciency in allocation of scarce resources. In both the West Bank and the Gaza Strip, UN Relief and Works Agency facilities are mainly concentrated in Palestinian refugee camps.
www.thelancet.com Vol 373 April 4, 2009

Series

Health services in the Gaza Strip have deteriorated rapidly since the political impasse between the Palestinian National Liberation Movement (known as Fatah) and the Islamic Resistance Movement (known as Hamas), and the Israeli and international boycott of Hamas,46 beginning in mid-2006 after the movements election victory.9 Secondary and tertiary care in the Gaza Strip are provided mainly by the Palestinian Ministry of Health, which is the only provider able to cope with the many cases and injuries related to the conict, indicating the burden of the deteriorating Palestinian National Authority health sector in the Gaza Strip.47 In June, 2007, Israel refused to allow travel outside the Gaza Strip for all patients referred to healthcare services abroad (282 cases),34 a policy indicative of Israels decision to impose collective restrictive measures against civilians in the Gaza Strip.48 Limited access to care has been documented in a series of reports prepared by the WHO oce in Jerusalem, showing emerging shortages of drugs, medical supplies, and equipment during 2006 and 2007.49 A recent WHO report documents 32 patients who died between October, 2007, and March, 2008, after being denied access to specialised treatment from outside the Gaza Strip.50 At the time of writing (December, 2008), some hospitals in the Gaza Strip lack basic health commodities, such as anaesthetics needed for surgery.51 A plan for human resource development was prepared in 2000,52 but its implementation has been poor. Although there are 26 physicians for every 1000 people in the Gaza Strip, there are only 18 in the West Bank.18 Furthermore, the ratio of allied health professionals to population is very low compared with that in other countries,36 and the skills of these workers are poorly developed and underused. Whereas Israel has 63 nurses for every 1000 people, the occupied Palestinian territory has only 17.18 Health-care services remain highly physician oriented,10 with doctors running many activities that could be done by nurses and community health workers at much lower costs.53,54 As a result of generalised unemployment24 and impoverishment25 since 2000, the Palestinian National Authority tried to soften the economic blow by putting more people on the government payroll,26,55 and has done so despite an increasingly unsustainable wage bill.26 Meanwhile, many workers have been hired to bolster political support.26 Between 1999 and 2006, public-sector employment grew by 60%,26 from fewer than 100 000 to 157 800, and increased further to 189 000 by mid-200756 under the Hamas-led government. An estimated 1 million Palestinians (workers and their dependants) depend on wages earned in public employment.57 Such costs are further increased when Palestinian clearance revenues are withheld by Israel and international nancial support is frozen, as happened between March, 2006, and July, 200726when Palestinian National Authority
www.thelancet.com Vol 373 April 4, 2009

Checkpoint Health facility (hospital or clinic) Access Dicult Very dicult Extremely dicult Separation wall Constructed Projected Under construction Tulkarm Nablus

Jenin

Tubas

Qalqiliya

Salt

Ramallah

Jericho

Jerusalem

Bethlehem

Dead Sea

Hebron

10 km

200

Figure: Distribution of and access to health facilities in the West Bank Planned barrier path based on Israeli Government map published Feb 20, 2005. Reproduced with permission from the UN Oce for the Coordination of Humanitarian Aairs.33

employees, including Ministry of Health sta, did not receive their usual salaries.18 Although nearly all health service sta during the Israeli military occupation of 1967 to 1994 were Palestinians, they had only token involvement in the decision-making process, with a consequent failure to promote Palestinian
1209

Series

West Bank Primary health-care centres Hospitals Hospital beds Population per primary health-care centre Beds per 1000 population Distribution of service use Health workers Distribution of health-care expenditures 525 54 2961 (591%) 4519 12

Gaza Strip 129 24 2053 (41%) 10 774 14

Ministry of Health* 416 (636%) 26 2936 (585%) 46% 13 057 35%

UNRWA Palestinian Private Occupied NGOs Palestinian territory 53 (81%) 1 63 (13%) 20% 1752 10% 185 (283%) 28 1582 (316%) 13% 7102 15% 23 433 (86%) 21% 7341 40% 654 78 5014 5752 13

NGOs=non-governmental organisations. UNRWA=UN Relief and Work Agency for Palestine Refugees. * Includes police medical services.

Table: Distribution of health-care facilities in the West Bank and the Gaza Strip according to provider18,40

leadership capacity.58 The available health education programmes cannot meet the needs of the health sector,18 and health professionals must therefore get training in schools outside the occupied Palestinian territory (mainly in Arab countries and in eastern Europe). The poor performance of the health-care system, and the obscure track of career progressioncommonly a function of personal connections and cronyismhas also led to loss of talented health professionals from Ministry of Health facilities to international and local non-governmental organisations, the private sector, and through migration outside the occupied Palestinian territory: in a study of 95 medical doctors, 29 had reported seriously considering emigration.59 Insucient monitoring and lack of supervision have allowed cronyism and corruption,60 a lack of commitment and interest, and erosion of public trust and satisfaction.12 An absence of requirement for continuing education activities52 compromises the quality of care and threatens patients safety. A health information system to support evidence-based policy formulation was devised; however, data collection, analysis, and reporting need further development. The second report in this Series10 shows how obtaining a reliable estimate of an indicator as basic as infant mortality is problematic. Many surveys are done and huge amounts of data are obtained that are improperly analysed. Barriers to promote a culture of evidence-based practices mainly relate to resistance to change and the feeling of a lack of ownership. The lack of eective partnership with those who are supposed to use the evidence, the almost absent dissemination of results, and the disengagement from the implementation process all culminate in loss of incentive to tackle change. The absence of proper nancial incentives (eg, in the form of results-based nancing) contributes to the entrenchment of current practices.
1210

The decit in the information needed for health-care management is a major diculty, restricting the capacity to plan and assess performance. Information is urgently needed about the most common problems (eg, non-communicable disease prevalence and complications) to assess cost-eectiveness of various programmes; nancial analysis is also needed to identify the treatments abroad with the largest budget shares; and patterns of use of non-governmental and Ministry of Health hospitals should be investigated to inform planning and improve eciency. Ministry of Health expenditure on drugs and disposables exceeds a fth of its actual expenditure (23% in 2005)2117% in Jordan and 25% Lebanon.61 The ministry implemented the rst Essential Drug List in 2000 and prepared a Drug Formulary in 2002. The eect of these initiatives has not been fully assessed.62 However, observational studies indicate that irrational prescribing is a continuing diculty,63 created by short consultation and dispensing times and absence of treatment guidelines.62 A recent review of medicines use in 6032 encounters with patients in selected non-governmental clinics showed that antibiotics (mostly wide-spectrum) accounted for 33% of medicines prescribed, followed by analgesic and anti-inammatory treatments (29%).62 Combination medicines (8%) and injections (16%) were also given in many cases. 77% of prescriptions were of locally produced medicines, which typically cost less than similar drugs of Israeli or international origin (eg, the average price of co-amoxiclav in 2004 was $525 if of local origin, $605 from Israel, and $685 from other sources).62 Data from nine eastern Mediterranean countries showed that medicines account for a high proportion of health expenditure (35% in Egypt and Jordan).64 Irrational prescribing and the higher prices of some prescribed medicines both contribute to the exaggerated burden on public expenditure and society as a whole. Investments in expensive medical technology do not also seem to have arisen from population-based requirements or from examining the cost-eectiveness of technologies.65 Health nancing consists of collecting revenues, pooling risks, and allocating resources to purchase services.66 The high proportion of GDP devoted to health67 was a direct result of both the high level of investment needed to rehabilitate a system that had been neglected between 1967 and 1994,58 and the low GDP. For example, between 1986 and 1989, while under the Israeli military government, the West Bank yearly public budget for health was kept at about $20 million, despite rising costs of living and population increases.68 After establishment of the Palestinian National Authority in 1994, the task of building a functioning health-care system has been supported by substantial donor assistance. Between 1994 and 2000, donors committed $353 million to the health sector and disbursed
www.thelancet.com Vol 373 April 4, 2009

Series

about half of that in actual assistance;69 by 2005, over 40 donors70 had contributed $10 billion to the Palestinian National Authority26 (most of which was donated by the European Union71). The inadequate coordination of funding sources has contributed to the diculty of even achieving simple objectives. Relief and emergency have repeatedly been the focus of health programmes, rather than long-term development, which has limited eectiveness, duplicated eorts, and eventually undone previous progress.10,11 This situation has perpetuated dependence on external funding, compromising sustainability and future self-suciency. Indeed, after the recent Palestinian National Authority nancial crisis, the Ministry of Health was not able to provide its essential operational budget,18 and most activities in the Medium Term Development Plan 2006200872 did not start.18 The Ministry of Health budget accounts for about 10% of the overall budget of the Palestinian National Authority.67 However, the nances of the ministry have been precarious since 2001, both because of a steep decline in health insurance revenues, and an increase in the number of people insured. The increase in insurance coverage was partly due to the decree issued by the then president of the Palestinian National Authority, Yasser Arafat, that all Palestinian victims of the intifada would be covered by the health insurance scheme, without contributions. From 2000 to 2002, the number of households covered by the new free insurance scheme increased by 205 430,21 whereas revenue from premiums fell from $295 million to $220 million.6 All revenues collected by the Ministry of Health are transferred to the Ministry of Finance, including any user fees and insurance co-payments; this practice undermines the potential incentive for improving the quality of care and providing some nancial exibility at a facility level.73 Indeed, the centralised nancial management of the Ministry of Health at the Ministry of Finance restricts the capacity of the directors of hospitals and health centres to exercise control over budgeting and stang.74 The Palestinian National Authority health insurance system covered 56% of Palestinian households in 2005, 55% of which were insured without contributions.21 The present fragmentation of the health nancing system makes risk pooling a real challenge and compromises horizontal equity (ie, the requirement that people with equal ability to pay make equal payments).75 Although the Palestinian National Authority compulsory health insurance is for public-sector employees only and insurance premiums do not adequately accord with ability to pay, the scheme does progressively assist poor people.13 Unlike most nancing plans elsewhere, the authoritys health-insurance scheme requires patients to pay for diagnostic services and drugs but not for consultation costs. This situation is contrary to the idea that patients are in control of the decision to consult doctors but exercise little control over diagnostic services used.36
www.thelancet.com Vol 373 April 4, 2009

In a recent survey of national health expenditure,15 households reported spending an average of 40% (range 961%) of their own out-of-pocket resources on health.6 Out-of-pocket payments are clearly regressive (ie, the proportional cost of payments decreases as ability to pay increases) and remain the most inequitable way of nancing health care;76 they are also unlikely to generate sucient revenue to support adequate services.77 The current structure of out-of-pocket payments in the occupied Palestinian territory has almost no pricediscrimination policies to account for dierences in ability to pay.13 In 2005, the total operating expenditure of the Ministry of Health was $1396 million (compared with $1003 million in 2000), of which 524% was spent on salaries, 226% on drugs and supplies, 157% on referrals outside the Ministrys facilities (excluding cases referred by the Palestininan National Authority Cabinet), and 93% on other operating costs.18 Scarce resources have been wasted by health-care provisions that were not cost-eective, with inappropriate focus on high-technology interventions and tertiary health-care services, which were mainly provided outside the area. The total cost of treatment outside the Ministry of Health (inside and outside the occupied Palestinian territory) was $581 million in 2004 and $596 million in 2005, constituting 460% and 427% of the ministrys expenditure, respectively.21 In 2005, more than $40 million was spent on services provided outside the occupied Palestinian territory, mainly in Jordan, Egypt, and Israel.21 In 2006, the number of patients referred elsewhere fell by 279%, lowering the costs of such treatment by 37%.78 The high expenditure on treatment outside the occupied Palestinian territory leads to loss of benets to the Palestinian economy.78 The World Health Report 2000 denes stewardship as the careful and responsible management of the wellbeing of the population,79 and refers to the responsibility of the state for the welfare of its population.80 Under Israeli military occupation, severe budgetary restrictions, underdevelopment58 and de-developmentthe destruction of the development process81became key features, including marginalisation of government health services and ad hoc dependence on Israeli medical services.68 Since 1994, the Ministry of Health has endeavoured to establish the role of the health sector, ideally through promotion of accessible, aordable, and sustainable health care of good quality and cost-eectiveness.82 The ministry is the steward of the system and ve of its six main stated functions relate to leadership and regulation.82 Since then, in addition to increasing the number of primary health-care centres and hospital beds and promotion of the Palestinian National Authoritys health insurance scheme, the ministry introduced new programmes (eg, reproductive health10 and dental care) and upgraded old ones (eg, school
1211

Series

health83 and chronic diseases11). The ministry is also responsible for licensing health professionals and institutions. Present practice, however, takes into account several input criteria in the licensing process with almost no consideration given to result and performance indicators. The prevailing circumstances restrict the capacity of the ministry to oversee the entire system, collectively plan its development, and exercise its regulatory role eectively. A health plan for the occupied Palestinian territory84 was developed with input from many stakeholders before the ministry was established. Since then, and before the third national health plan,18 non-ministry stakeholders were not eectively involved in the planning process, with the result that there has been no overall development policy around which national and provider-specic policies could be developed. Although many of the objectives in successive national strategic health plans have been clear and restricted, few have had target completion times, or adequate budget preparation and prioritisation6 (this is being attempted in the latest plan18). Furthermore, there have been no regular reviews and updates of stated objectives, taking account of achievements and changing circumstances.6 Although preparation began in 2003, the third health plan18 was only nalised in 2008.

System building under military occupation

Several attempts to build a health-care system for the occupied Palestinian territory have been made in recent years, with some advances being achieved against the odds.58 However, despite the substantial amount of money injected into the system6 and the two concluded national health plans,82,85 systemic goals remain far from met.9,12,13 This failure is mainly due to three inter-related factorsendogenous Palestinian features, donors policies, and political havocthat compromise the WHO building blocks. The Palestinian National Authority is expected to perform as the government of a state86 while lacking control over its borders, basic resources, and many of the social determinants of health. The absence of a long-term Palestinian development agenda focusing on sustainable and equitable growth26 has compromised any strategic planning. The extensive overlap of specialisation and the duplication of functions result from inadequate delineation of responsibilities, which impedes formulation, planning and prioritisation, implementation, and assessment of policies.87 Internal mismanagement has promoted divides within the divide, with each stakeholder seeking its own nancial survival.60 The highly centralised nancial and stang systems and the lack of will to provide value for money have impeded development. Vague institutional arrangements have hindered the establishment of a proper governance system characterised by transparency, separation of powers, and the rule of law.55
1212

Despite the important role of external funding in alleviating short-term eects of a socioeconomic crisis, the existing confusion in the system is compounded by the multiplicity of donors, who can have conicting agendas and be poorly coordinated.58 Notwithstanding several initiatives for joint programmes,88 aid in the occupied Palestinian territory has repeatedly been reactive and has not always encouraged institutionbuilding or created incentives for reform.70 Many donations are based on bilateral deals that suit the donors political needs and preferences as much as Palestinians requirements.70 A 2007 document26 presented to the Ad Hoc Liaison Committee89 established to support the Middle East peace process recognises that donations remained fragmented and focused on bilateral arrangements with donors based on short-term political positions rather than a collective, longer-term view on broader economic and governance fundamentals.26 A recent report called for even more donations to enhance growth and enable development.90 The political instability of the Palestinian National Authority, with frequent ministerial changes (six ministers of health have been nominated in the past 3 years), has contributed to system instability. In the local context, many positions are lled on the basis of political favouritism55 and ministerial changes are commonly accompanied by changes in mid-level and high-level managerssometimes by additional recruitment.56 The factors that hinder health system development are not unique to the occupied Palestinian territory, but they are exaggerated and perpetuated under the oppressive conditions of the Israeli military occupation.60 Furthermore, occupation creates some of the diculties.91 Occupation policies of separation, isolation, and segregation have created uncertainty, raised transaction costs, and shrunk markets, resulting in critical constraints on the survival of the Palestinian economy as a whole.26 A recent World Bank report states that: growth rates will depend critically on the commitment of the international community to ll the total scal gap[nevertheless]Even with full funding but no relaxation in the closure regime, growth will be slightly negative.90 Admittedly, as Ajluni92 has said, imagining a rational system of planning and nancing is dicult when Israeli policy has greatly damaged infrastructure and impoverished the population.92 The intensied siege and closure of the Gaza Strip has complicated already dicult reform eorts;18 and the uncertainty about future developments, imposed by a fruitless peace process, aggravate the situation further.

A way forward
Considering that Israel has never dened its borders,93 the feasibility of steps to improve the health system in the occupied Palestinian territory will depend on future
www.thelancet.com Vol 373 April 4, 2009

Series

political developments and border denitions, and the commitment on the part of the Palestinian society and the Palestinian National Authority to eect change. Although a best-case scenario would include establishment of a sovereign Palestinian State on all of the Palestinian territory occupied in 1967 (in accordance with UN resolution 24294), other possible scenarios can also be envisaged. First, an enduring status quo, with continued construction of Israeli settlements and the separation wall in the West Bank (both deemed illegal under international law95,96), continued separation of the Palestinian Arab East Jerusalem from the West Bank and the Gaza Strip under Israeli military occupation, and continued political impasse between Fatah and Hamas.9 Second, a worst-case scenario would include a worsening political situation leading to the collapse or dissolution of the Palestinian National Authority, as a result of the persistent failure of peace talksa situation now seen as plausible (or even imminent) by high level ocials in the authority.97 Dissolution of the authority would result in one of three possible developments: Israel resuming full responsibility, as a signatory of the Fourth Geneva Convention,98 for the Palestinian population, which it now controls de facto; a return to the pre-1967 arrangements, with Jordan resuming control of what remains of the West Bank and Egypt administering the Gaza Strip;99 or the systematic, illegal,98 expulsion of Palestinians to neighbouring Arab countries, as repeatedly suggested by some former100 and current101 Israeli leaders. Should a Palestinian governing body continue to exist, the most important feature of any successful initiative toward building an eective, ecient, and equitable health-care system will be eective stewardship, by which the Palestinian Ministry of Health becomes empowered and has the capacity to oversee and steer the entire system. A clear vision is needed of the regulations to be put in place, with frameworks for monitoring and evaluation being essential. Commitment from stakeholders other than the ministry is also important stewardship after all is about collective rather than individual responsibility.80 The high cost of treatment abroad might be countered by persuading partner sectors to provide services the ministry cannot aord. Here, the private sector can play an important part.78 Moreover, focus should be given to stewardship for health, rather than just for health care, calling for intersectoral collaborations. A clear policy for human resources for health is needed. Eorts to form, upgrade, and integrate individual capacities would not only enable future development plans to succeed102 but also help build experience and condence and introduce a momentum to promote change. As one observer of the situation in the occupied Palestinian territory put it, it is the strong
www.thelancet.com Vol 373 April 4, 2009

individual capacities against severe institutional weaknesses that has enabled a system of sorts to survive.103 Creation of an eective human-resources capacity should start with revision of the available national plan for human resources52 to identify needs in terms of health professionals and education and training programmes, as related to predened strategic objectives. Rather than merely being providers of academic services, universities should build the required capacities, for example, through continuing education ventures. Strengthening of monitoring and supervision, accompanied by a system of rewards and sanctions, seems crucial to boost motivation, to halt the brain drain, and to enhance eectiveness and future development. Eorts should be made to remedy malfunctioning schemes of health-care nancing. Concrete steps are needed towards the institutionalisation of a genuine social insurance scheme with a view to universal coverage. The current health insurance scheme should become a sovereign and accountable legal entity, functioning under collective ministerial supervision, with control over its own revenues, which can be used to purchase services with appropriate methods of nancing.104 Community-based health nancing schemes and the current Palestinian National Authority health insurance together could form a nucleus to enable ecient and equitable resource mobilisation105 and to introduce the change gradually, taking into consideration prevailing political, economic, and social conditions. Recommendations for health-sector reform suggest that payment by capitation for primary care and per admission for inpatient care might improve cost-eectiveness and increase equity.76 In the local context, evidence suggests that patients are willing to pay to benet from improved essential quality attributes,106 with amounts varying according to the extent of improvement and patients abilities to pay.107 Such information could be used to help identify areas for improvement and inform the pricing structure in the adopted nancing scheme. Decisions should be evidence based, and for this to be possible, an accurate and continuously updated health information system is needed. Culturally sensitive evidence is needed, followed by the development and implementation of national protocols and standard operational procedures. However, only by involvement of those to whom the evidence is directed (eg, healthcare providers and decision makers) and follow-up of the implementation process would practices eectively change to help full the three systemic goals improving health, responding to expectations, and enhancing risk protectionin the most cost-eective way. Quality of care should be at the core of any endeavouracknowledging that improvement in quality does not always mean higher cost, but it does demonstrate political commitment (panel).
1213

Series

Panel: Six WHO building blocks for health systems and the way forward for the occupied Palestine territory Service delivery Promote primary health care as the eective backbone of the health-care system by increasing investment in primary centres and public-health activities (eg, preventive programmes for chronic diseases) Integrate and coordinate all providers and human-resources activities, with clear division of roles and tasks Workforce Revise the available plan for human resources and its concordance with established policies and plans and prevailing gaps Assess available training programmes and activities in terms of their quality and appropriateness Build human capacities in planning, nancing, and provision of health care Strengthen monitoring and supervision Develop a results-based system of rewards and sanctions, with transparent tracks of career progression Information Upgrade the current health-information system to provide the information needed for all levels of clinical and administrative decision-making processes Develop national clinical management and administrative protocols, while involving those to whom the evidence is addressed in the process Promote a culture of evidence-based decision making at all levels of health-care planning and provision through hands-on training and intensive follow-up and supervision Medical products and technology Promote rational use of drugs and eective drug management to increase accessibility and avoid wastage of scarce resources Revise practices of purchasing, prescribing, and dispensing to focus on the most cost-eective medications and technology Financing Transform the current Palestinian National Authority health insurance scheme into a sovereign and accountable legal entity with control over its own resources Promote community-based health nancing to cover the health-care costs of various categories of the population Integrate a system of payment by capitation for primary health care and per admission for inpatient care Work towards institutionalisation of a genuine universal scheme of social insurance Establish a single treasury account for the donors to reduce duplication and wastage of resources Stewardship Empower the Ministry of Health through appropriate regulations and enhance capacity of planning and supervision Collectively redene a vision for the health system Enhance intersectoral collaboration at all levels of planning, nancing, and provision End cronyism and growth of public employment

The view of health in Palestinian Public Health Law as a fundamental human right should be integral to any eorts to establish a Palestinian State, while acknowledging limitations of resource availability and the need for self-suciency. Promotion of primary health care and integration of providers activities and available human resources (medical, paramedical, and
1214

non-medical) will ensure the most benets for the most people. Preventive programmes are needed to enable favourable long-term health outcomes at low costs (eg, tobacco control and diet promotion eorts that reduce the burden of chronic diseases11). The Palestinian Ministry of Healths ability to set priorities that address individual preferences and to negotiate with donors should be improved to help avoid conicting agendas and promote the welfare of the population. Under the best-case scenario, the Palestinian Ministry of Health can choose to limit its role to being the steward of the system, while providing a basket of core services (including public-health activities), and the funding needed to cater for the needs of specic categories of the population (eg, vulnerable groups). Services could be purchased by an independent insurance fund that would raise money through an appropriate insurance-based risk-pooling mechanism to gradually create a selfsucient, ecient, and equitable health-care system. Under the worst-case scenario, the focus has to remain on emergency relief, with attempts made to pursue human capacity development to enable the change, if circumstances allow. Finally, the scenario of the continuing status quo means consideration should be given to coordinating and harmonising the eorts of donors and providers to avoid wastage of scarce resources. Such synergy can be established by ending cronyism and ination in public employment, establishing a single treasury account for the donors, while striking the balance between recruited sta and task allocation by putting competent professionals in the right places, by decentralisation of decision making, by promoting team work, and by involving the community. A new opportunity to improve Palestinians quality of life and increase national prosperity108 emerged at the Paris conference in December, 2007.90 Delegates decided to allocate a substantial sum of money to the Palestinian National Authority, including about $120 million to be raised for the health sector between 2008 and 2010. The best allocation and use of these new funds will plant the seeds of genuine and sustainable development. But health systems do not evolve in a vacuum. For continued development, social determinants have to be addressed. In the case of the occupied Palestinian territory, the occupation has to end. For a long time, Palestinians have declared we are here to stay. What Palestinians want was summarised in the words of a woman in a recent article that appeared in The New York Review of Books: We want to live in peace and dignity our suering will not end without ending the occupation.109
Contributors AM wrote the rst draft, all coauthors helped revise subsequent drafts before submission.

www.thelancet.com Vol 373 April 4, 2009

Series

Conict of interest statement We declare that we have no conict of interest. Acknowledgments We thank the Lancet Palestine Steering Group (Iain Chalmers, Rita Giacaman, Jennifer Leaning, Harry Shannon, and Huda Zurayk) for having read, discussed and commented on several drafts of this report; Firas Raad, Graham Watt, Sawsan Batato, and Karl Sabbagh for their valuable comments; Medical Aid for Palestinians UK and Oslo University Medical School for their nancial contributions that made the workshops related to this Series possible; and the anonymous reviewers, whose comments improved this report substantially. References 1 UN High Commissioner for Human Rights. The right of everyone to the enjoyment of the highest attainable standard of physical and mental health: Commission on Human Rights resolution 2002/31, New York: UN, 2002. 2 McKee M. Measuring the eciency of health systems. The world health report sets the agenda, but theres still a long way to go. BMJ 2001; 323: 29596. 3 Chalmers I. Human rights and professional responsibilities in health care and education. Eur J Dent Educ 2008; 12: 1116. 4 Commission on Macroeconomics and Health. Macroeconomics and health: investing in health for economic development. Geneva: WHO, 2001. 5 UN. Question of Palestine: history. http://www.un.org/Depts/dpa/ ngo/history.html (accessed March 20, 2008). 6 DFID. West Bank and Gaza health sector expenditure review, 2006. London: Department for International Development, 2006. 7 Roberts M, Hsiao W, Berman P, Reich R. Getting health reform right: a guide to improving performance and equity. Oxford: Oxford University Press, 2003. 8 WHO. Everybody business: strengthening health systems to improve health outcomes: WHOs framework for action. Geneva: World Health Organization, 2007. 9 Giacaman R, Khatib R, Shabaneh L, et al. Health status and health services in the occupied Palestinian territory. Lancet 2009; published online March 5. DOI:10.1016/S0140-6736(09)60107-0. 10 Abdul Rahim HF, Wick L, Halileh S, et al. Maternal and child health in the occupied Palestinian territory. Lancet 2009; published online March 5. DOI:10.1016/S0140-6736(09)60108-2. 11 Husseini A, Abu-Rmeileh NME, Mikki N, et al. Cardiovascular diseases, diabetes mellitus, and cancer in the occupied Palestinian territory. Lancet 2009; published online March 5. DOI:10.1016/S0140-6736(09)60109-4. 12 World Bank, Bisan Center for Research and Development. The role and performance of Palestinian NGOs: in health, education and agriculture. Washington, DC: The World Bank. 13 Abu-Zaineh M, Mataria A, Luchini S, Moatti J-P. Equity in health care nancing in the Palestinian Context: the value-added of the disaggregate approach. Soc Sci Med 2008; 66: 230820. 14 Mataria A, Raad F. Analyzing health equity in the West Bank and Gaza. Washington, DC: The World Bank, 2009. 15 PCBS. Households health expenditure survey 2004. Ramallah (OPT). Palestinian Central Bureau of Statistics, 2004. 16 Batniji R, Rabaia Y, Nguyen-Gillham V, et al. Health as human security in the occupied Palestinian territory. Lancet 2009; published online March 5. DOI:10.1016/S0140-6736(09)60110-0. 17 Palestinian Central Bureau of Statistics. Population, housing and establishment census2007. http://www.pcbs.gov.ps/Portals/_ pcbs/PressRelease/census2007_e.pdf (accessed Jan 14. 2009). 18 Ministry of Health. National strategic health plan: medium term development plan (2008-2010). Palestine: Palestinian National Authority, Ministry of Health, Health Planning Unit, 2008. 19 UNRWA. The United Nations Relief and Works Agency for Palestine Refugees in the near east guide. Vienna: UNRWA, Public Relations Department, 1995. 20 Abed Y. Health sector review: a summary report. Jerusalem: The Italian Cooperation, 2007. 21 MoH-PHIC. Health status in Palestine: Annual Report 2005. Ramallah (OPT): Ministry of HealthHealth Management Information System, 2006. 22 WHO. Health action in crises: West Bank and Gaza Strip. Geneva: World Health Organization, 2005.

23

24 25

26

27

28

29 30

31

32

33

34

35

36

37 38

39

40

41

42

43

44 45 46

The World Bank. West Bank and Gaza: economic developments and prospects, March 2008. http://go.worldbank.org/A6Y2KDMCJ0 (accessed Aug 21, 2008). PCBS. Palestine in gures 2007. Ramallah: Palestinian Central Bureau of Statistics, 2008. PCBS. Poverty and living conditions in the Palestinian territory, 2007. http://www.pcbs.gov.ps/Portals/_pcbs/PressRelease/ poverty_ee.pdf (accessed Aug 18, 2008). World-Bank. Two Years after London: Restarting Palestinian Economic Recovery Economic Monitoring Report to the Ad Hoc Liaison Committee. http://conictblotter.com/les/ WorldBankReportSEP07.pdf (accessed April 2, 2008). UN. Commission on Human Rights, 60th session: economic, social and cultural rightsthe right to food http://domino. un.org/UNISPAL.NSF/f45643a78fcba719852560f6005987ad/bc 2cc84ae9bc85256e2b00685371!OpenDocument (accessed Aug 19, 2008). Mataria A, Khatib R, Daoud A. Patients ow and referral system at PMRS clinics in the West Bank: status quo, variations and determinants. Ramallah: Institute of Community and Public Health, 2007. PCBS. Access to health services survey2003: main ndings. Ramallah: Palestinian Central Bureau of Statistics, 2004. OCHA. West Bank closure count and analysis occupied Palestinian territory, September 2006. http://www.ochaopt.org/ ?module=displaysection&section_id=103&format=html (accessed Nov 30, 2007). Habib I. A wall in its midst: the separation barrier and its impact on the right to health and on Palestinian hospitals in East Jerusalem. Tel Aviv: Physicians for Human Rights, 2005. Jubran J. Health and segregation II: the impact of the Israeli segregation wall on access to health care services. Ramallah: Health Development, Information, and Policy Institute, 2005. UN Oce for the Coordination of Humanitarian Aairs. Health facilities map-2. http://www.ochaopt.org/documents/WB_Health_ CAP2006.pdf. Oce for the Coordination of Humanitarian Aairs. The humanitarian monitoroccupied Palestinian territory. Oce for the Coordination of Humanitarian Aairs, 2007. Daoud A, Khatib R. Capacity building on rational use of medicine: knowledge and practice aecting the mobile teams. Ramallah: Institute of Community and Public Health, Birzeit University, 2007. Schoenbaum M, A A, Deckelbaum R. Strengthening the Palestinian health system. Santa Monica: RAND Corporation, 2005. Barghouthi M, Lennock J. Health in Palestine: potential and challenges. Ramallah: MAS Discussion Papers, 1997. Hana S, Tabar L. The Intifada and the aid industry: the impact of the new liberal agenda on the Palestinian NGOs. Comp Studies South Asia, Africa, Middle East 2003; 23: 20514. WHO. Palestinian health status. http://www.who.int/hac/crises/ international/wbgs/oPt_MoH_Palestinian_Health_Status_ June2006.pdf. (Aug 19, 2008). Palestinian Central Bureau of Statistics. Health care providers and beneciaries survey2005: main ndings. Ramallah: Palestinian Central Bureau of Statistics, 2006. Mataria A, Khoury P. Public policies to enhance private sector competitiveness in providing tertiary health care in palestine: assessment and recommendations. Ramallah: Palestine Economic Policy Research Institute, 2008. WHO. Health action in crises: West Bank and Gaza Strip. http://www.who.int/hac/crises/international/wbgs/OPT_Apr05. pdf (accessed Aug 21, 2008). Mataria A, Giacaman R, Khatib R, Moatti JP. Impoverishment and patients willingness and ability to pay for improving the quality of health care in Palestine: an assessment using the contingent valuation method. Health Policy 2006; 75: 31228. Sen A. Development as freedom. New York: Alfred A Knopf, 1999. Sen A. Inequality Re-examined. Oxford: Clarendon Press, 1992. Devi S. Hospitals in Gaza struggling after factional ghting. Lancet 2007; 369: 214748.

www.thelancet.com Vol 373 April 4, 2009

1215

Series

47

48

49

50

51

52

53

54

55

56

57

58

59

60 61

62

63 64

65 66

67

68

The Gaza Strip: A humanitarian implosion. http://news.bbc. co.uk/1/shared/bsp/hi/pdfs/06_03_08_gaza.pdf (accessed March 16, 2008). PHR. Physicians for human rightsIsrael update. http://www.mapuk.org/images/content/File/Physicians%20for%20Human%20Rig hts.pdf (accessed Jan 16, 2007). WHO. Monitoring the Health Sector in the occupied Palestinian territory, series reports. http://www.emro.who.int/palestine/index. asp?page=library&option=monitoring.WHO_monthly_monitoring (accessed Jan 7, 2008). WHO. Access to Health Services for Palestinian people: case studies of ve patients in critical conditions who died while waiting to exit the Gaza Strip. West Bank and Gaza: WHO, 2008. WHO. Health sector surveillance indicators: monitoring the health sector in the occupied Palestinian territory. http://www.emro.who. int/palestine/index.asp?page=library&option=monitoring.WHO_ monthly_monitoring (accessed Jan 7, 2008). MoEHE, MoH, WA. National plan for human resource development and education in health. Ramallah: Ministry of Education and Higher Education, Ministry of Health, Welfare Association, 2001. Townsend J, Wolke D, Hayes J, et al. Routine examination of the newborn: the EMREN studyevaluation of an extension of the midwife role including a randomised controlled trial of appropriately trained midwives and paediatric senior house ocers. Health Technol Assess 2004; 8: 1100. ICPH-BZU. Evaluation of the UNFPA reproductive health sub programoccupied Palestinian territory (oPt): 2001-2005. Ramallah: Institute of Community and Public Health, Birzeit University, 2006. Said N, Badawi W. Public administration in the West Bank & Gaza: obstacles and opportunities. http://unpan1.un.org/intradoc/ groups/public/documents/un/unpan016109.pdf (accessed March 28, 2008). Yaghi M. Fayads reform plan: diculties and political implications. http://www.washingtoninstitute.org/templateC05.php?CID=2729 (accessed April 3, 2008). The Portland Trust. Debt rises in majority of Palestinian households. Palestinian Econ Bull 2007; 10. http://www. portlandtrust.org/Bulletin%20Issue%2010%20July%202007.pdf (accessed Jan 14, 2009). Giacaman R, Abdul-Rahim HF, Wick L. Health sector reform in the occupied Palestinian territories (OPT): targeting the forest or the trees? Health Policy Plan 2003; 18: 5967. Mataria A, Abu-Hantash I, Amer W. The brain drain of the Palestinian society: with an exploratory study of the health and higher education sectors. Ramallah: Palestine Economic Policy Research Institute (MAS), 2008. Halevi I. Self-government, democracy, and mismanagement under the palestinian authority. J Palestine Studies 1998; 27: 3548. Khatib R, Mirza Z, Mataria A. Access to essential medicines in the Arab world: impediments and areas for improvement. In: Nuwayhid I, Khawaya M, Jabbour S, Giacaman R, eds. Public health in the Arab world. Cambridge: Cambridge University Press (in press). Khatib R, Daoud A, Abu-Rmeileh NM, Mataria A, McCaig D. Medicine utilisation review in selected non-governmental organisations primary healthcare clinics in the West Bank in Palestine. Pharmacoepidemiol Drug Saf 2008; 17: 112330. PSF. Medical and pharmaceutical situation inside the West Bank in primary health care. Ramallah: Pharmaciens Sans Frontier, 2003. WHO-EMRO. Technical discussion on: medicine prices and access to medicines in the Eastern Mediterranean Region. Cairo: World Health Organization, 2007. Garner P. Expensive medical technologyconsultancy report. Boston: Harvard Institute for International Development, 1997. Carrin G, James C. Key performance indicators for the implementation of social health insurance. Appl Health Econ Health Policy 2005; 4: 1522. Hamdan M, Defever M, Abdeen Z. Organizing health care within political turmoil: the Palestinian case. Int J Health Plann Manage 2003; 18: 6387. Giacaman R. Health conditions and services in the West Bank and Gaza Strip: United nations Conference on Trade and Development. UNCTAD/ECDC/SEU/3, 1994.

69

70

71

72 73 74

75

76

77

78

79 80

81 82 83

84

85

86 87

88 89

90

91

92

Government of Japan, The World Bank. Aid eectiveness in the West Bank and Gaza. Jerusalem: The Secretariat of the Ad Hoc Liaison Committee, 2000. Conict Blotter. Palestinian economic news, September 18th, 2007. http://conictblotter.com/2007/09/18/palestinian-economic-news/ (accessed April 02, 08). Guttman N. Bush, Congress threaten to cut PA aid. http://www. jpost.com/servlet/Satellite?cid=1137605934575&pagename=JPost%2 FJPArticle%FShowFull] (accessed April 03, 2008). Ministry of Planning. Medium term development plan. Ramallah: Palestinian Authority, Ministry of Planning, 200608, 2005. World-Bank. Financing health services in development countries: an agenda for reform. Washington, DC: World Bank, 1987. Mataria A, Khatib R, Giacaman R. An assessment of the training needs for general, administrative and nursing hospital directors: the case of the public hospital in the West Bank and Gaza Strip. Ramallah: Institute of Community and Public Health, 2004. Abu-Zaineh M. Contribution to the analysis of health equity: measurement and explanation of inequality and inequity in the health care sector in the context of developing countries: the Palestinian case. PhD thesis. Faculty of Economics Sciences and Management, University of the Mediterranean, 2008. Gottret P, Schieber G. Health nancing revisited: a practitioners guide. Washington: The International Bank for Reconstruction and Development/The World Bank, 2006. Pauly MV, Zweifel P, Scheer RM, Preker AS, Bassett M. Private health insurance in developing countries. Health A (Millwood) 2006; 25: 36979. Mataria A, Khouri P. Public policies to enhance private sector competitiveness in providing tertiary health care in Palestine: assessment and recommendations. Ramallah: Palestine Economic Policy Research Institute (MAS), 2008. WHO. World health report 2000: health systems: improving performance. Geneva: World Health Organization, 2000. Hunter D, Shishkin S, Taroni F. Steering the purchasers: the stewardship role of government. In: Robinson R, ed. Eective Purchasing for Health Gain. Milton Keynes: Open University Press and European Observatory on Health Care Systems, 2004. Roy S. De-development revisited: Palestinian economy and society since Oslo. J Palestine Studies 1999; 28: 6482. NSHP. National strategic health plan (19992003). Gaza: Palestinian National Authority, 1999. Ziara S, El-Khateeb W, Skeik M, Abu-Arkoub O, Accorsi S, Cogo C. Health prole of the 1st grade school-aged children in Palestine: an account of a National School Health Programme. Ramallah: Palestinian National Authority Ministry of Health, 1997. The Palestinian Planning and Research Centre. National Strategic Health Plan (19911993). Jerusalem: The Palestinian Planning and Research Centre, 2001. NHP. National health plan for the Palestinian people: objectives and strategies. Jerusalem: The Palestinian Planning and Research Centre, 1994. Sayigh Y. Inducing a failed state in Palestine. Survival 2007; 49: 739. Sayigh Y, Shikaki K. Strengthening Palestinian public institutions. http://www.cfr.org/content/publications/attachments/palinstfull. pdf (accessed Sept 26, 2008). UN. Second multi-year funding framework, 2004-2007. www.undp. org/execbrd/word/dp03-32.doc (accessed Sept 26, 2008). Partners in Peace. The local aid coordination committee (LACC) for development assistance in the West Bank and Gaza Strip: http://www.arts.mcgill.ca/MEPP/unsco/pip/intro_les/I_2.html (accessed Sept 30, 2008). The World Bank. Investing in Palestinian economic reform and development: report for the pledging conference. http://www. prospectsforpeace.com/Resources/Statements/World_BankPalestinian_Economic_Reform_and_Development.pdf (accessed April 2, 2008). IRIN. ISRAEL-OPT: Palestinian economy to be dependent on aid in medium term. http://www.irinnews.org/Report. aspx?ReportId=75885 (accessed April 2, 2008). Ajluni S. The Palestinian Economy and the Second Intifada. J Palestine Studies 2003; 32: 6473.

1216

www.thelancet.com Vol 373 April 4, 2009

Series

93

Various authors. Swans commentaryinstructive quotations behind the Israeli propaganda. Accessed at: http://www.swans.com/ library/dossiers/ipquotes.html (accessed March 23, 08). 94 The UN Security Council. Resolution 242 (1967) of 22 November 1967. http://daccessdds.un.org/doc/RESOLUTION/ GEN/NR0/240/94/IMG/NR024094.pdf?OpenElement (accessed March 20, 2008). 95 The UN Security Council. Resolution 456(1980) of 1 March 1980. http://domino.un.org/unispal.NSF/f45643a78fcba719852560 f6005987ad/5aa254a1c8f8b1cb852560e50075d7d5!OpenDocument (accessed March 25, 2008). 96 ICJ. IsraelIsraels separation barrier: challenges to the rule of law and human rights: part III and IV. http://www.icj.org/news. php3?id_article=3411&lang=en (accessed March 20, 2008). 97 PIC. Erekat: PA is on the verge of collapse and Fatah is losing popularity. Ramallah: Palestinian Information Centre, 2008. 98 Oce of the High Commissioner for Human Rights. Geneva convention relative to the Protection of civilian persons in time of war. Geneva: United Nations, 1949. 99 Eshel D. Inner strife splits Gaza from West Bank: defense update. http://www.defense-update.com/newscast/1206/analysis/ analysis-171206.htm (accessed March 31, 2008). 100 Joseph Weitz. A solution to the refugee problem: an Israeli state with a small Arab minority. http://www.swans.com/library/dossiers/ ipquotes.html (accessed March 23, 2008).

101 Moledet. Israels foremost right-wing political party Moledet website. http://www.moledet.org.il/english/ (accessed March 25, 2008). 102 Adjibolosoo S. International perspectives on the human factor in economic development, 1998. Westport, CT: Praeger, 1998. 103 Pavignani E. Managing aid in the Palestinian health sector: an exploration. 2007. 104 Hsiao WC. Why is a systemic view of health nancing necessary? Health A (Millwood) 2007; 26: 95061. 105 Preker A, Carrin G, eds. Health Financing for poor people: resource mobilization and risk sharing. Washington: The International Bank for Reconstruction and Development/ The World Bank, 2004. 106 Mataria A, Donaldson C, Luchini S, Moatti JP. A stated preference approach to assessing health care-quality improvements in Palestine: from theoretical validity to policy implications. J Health Econ 2004; 23: 1285311. 107 Mataria A, Luchini S, Daoud Y, Moatti JP. Demand assessment and price-elasticity estimation of quality-improved primary health care in Palestine: a contribution from the contingent valuation method. Health Econ 2007; 16: 105168. 108 Palestinian National Authority. Building a Palestinian state: toward peace and prosperity. http://imeu.net/engine/uploads/pna-fullreport.pdf (accessed Jan 10, 2009). 109 Horton R. Palestinians: the crisis in medical care. New York Rev Books 2007; 54: 2224.

www.thelancet.com Vol 373 April 4, 2009

1217

Viewpoint

The danger of in-kind drug donations to the Global Fund

Brook K Baker, Eva Ombaka
Lancet 2009; 373: 121821 Published Online October 9, 2008 DOI:10.1016/S01406736(08)61487-7 Health GAP Northeastern University School of Law, Program on Human Rights and the Global Economy, Boston, MA, USA (Prof B K Baker JD); and Ecumenical Pharmaceutical Network, Kenya, Africa (E Ombaka PhD) Correspondence to: Prof Brook B Baker, Health GAP Northeastern University, School of Law Program on Human Rights and the Global Economy, 400 Huntington Avenue, Boston, MA 02115, USA B.Baker@neu.edu

In-kind drug donations to the Global Fund to ght AIDS, tuberculosis, and malaria (GFATM) could greatly distort market incentives for entry of generic drugs, and thus delay the development of a more competitive market for priority drugs. Donations can be detrimental to market incentives for generic competition because they reduce the size of the residual market for a particular drug, create uncertainty about future market eects of donations of other drugs, change the riskbenet ratio with respect to patent-related issues, registration barriers, and costs of negotiating a distribution system, and reduce the market-pull advantage of an identiable, sizeable, and sustained source of secure nancing. Similarly, drug donations can distort choices about best treatment options by deterring use of eective xed-dose combinations and other improved formulations, tying government treatment choices to particular so-called free drugs, delaying or undermining governments motivation to select or modify treatment guidelines, aecting regulatory decisions (such as registration of generic equivalents), and complicating procurement and supply systems. The issue of in-kind donations of drugs has been discussed within GFATM since 2002, mainly at the insistence of the private sector and certain donors.1 Recently a joint steering committee of GFATM has commissioned a policy paper analysing in-kind donations, including potential market and sovereignty eects, for consideration at the Boards November, 2008, meeting. Civil society delegations have consistently and successfully opposed drug donations. We present key arguments against two underexplored aspects of drug donations to GFATM namely, their distortion of market incentives and their adverse eects on therapeutic options. Drug donations to GFATM are tempting because they lower commodity costs and could increase the value of private-sector contributions. However, these apparent advantages are more than oset by the fact that drug donations might have serious adverse eects on market incentives for entry of generic drugs, and thus delay the development of a competitive and low-priced market for medicines. Similarly, drug donations can distort therapeutic options for best treatment. Up to now, analyses of drug donationsfor example, WHO guidelines,2,3 and studies for tropical or neglected diseases46 and for disaster relief7,8have neglected market distortion eects. An early analysis showed that costs of donations were four times higher than other alternatives such as discount prices or cash donations, after taking taxes into account.9 WHO prefers cash donations in the aftermath of emergencies,10 and both AIDS-related and malaria-related donations have proven problematic because of transparency, market segmentation, and conditionality issues4 or cost-inecient,11 but not on the basis of any thorough market analysis.

In the absence of market studies, analysis of related evidence for entry of generic drugs is useful in general,1227 and of antiretroviral medicines in particular.4,2831 The main ndings of these studies are that the number of generic entrants: (1) aects the extent of price competition,12,13 and the presence of eight or more competitors is related to the greatest price reductions;13,29 (2) positively correlates with market size (volume of sales) and expected prots,13,14,16,18,22,24 and with time since generic entry;23,24 and (3) negatively correlates with the number of incumbent generic rms,13,18,22 the number of competing drugs in a therapeutic class,14,22 the duration of the exclusivity of a products brand name (which builds brand loyalty),25 the presence of an early-entry branded generic,19,20 and, paradoxically, of reference pricing in one study22 but not another.15 The fact that market size and expected prots are the main drivers of entry of generic drugs is not surprising. Similarly, if the brand-name company captures post-patent market share via a branded generic or low price drug, negative eects on entry of generic drugs and ultimately price, not surprisingly, take place. Both ndings show the probable eect of drug donations on the aggregated markets of GFATM recipients, because donations also capture market share, and thus demotivate generic entrants. The basic economic argument rests on the fact that generic companies assess market prospects, uncertainty, and risk on the basis of anticipated costs, expected returns, and a prediction about how many competitors will enter or remain in the market. In making economic forecasts about whether investment of scarce resources will be protable, generic companies must address four key issues discussed below. How much aggregate market demand will there be and will it be possible to sell in dierent markets or will the product be restricted to a few markets or market segments only? Aggregating demand from dierent developing countries, including middle-income and low-income countries, and public, non-governmental and faith-based organisations, and private sectors makes the market attractive for generic entrants. If donations split the market or target specic sectors, generic entrants would be left with a small and dicult market niche and a disadvantageous costbenet ratio.4 Alternatively, if the market is large, generic companies can produce at more-ecient economies of scale; if the market is large enough, several companies will enter, promoting price competition. How will demand grow (or shrink) over time in relation to competition from other medicines and other generic producers?
www.thelancet.com Vol 373 April 4, 2009

1218

Viewpoint

Uncertainty always exists in pharmaceutical markets because innovators might produce improved products or new therapeutic alternatives. Moreover, what makes the market attractive for one generic entrant makes it simultaneously attractive for many entrants, aecting protability. Because generic producers must recoup their investment costs over time, the inability to be certain about size, growth, duration, and competitiveness of the eventual market increases the risk of unrecoverable sunk costs or leads to higher prices at product launch. What is the riskbenet ratio of costs of research and development, patent and registration barriers, and product distribution systems for markets aected by donations? Generic companies face entry costs and other barriers that are proportionately higher in markets aected by donations than in those that are not. For example, generic companies ordinarily spend at least US$115 million sometimes much moreto reverse engineer and formulate a product,32 after which they must invest in production processes and capacity. Before producing commercially, generic companies must be concerned about patents in countries of production and use.33,34 Even the establishment of patent status is often dicult,33 and when patents exist, generic producers must pursue compulsory licences,35 which is costly, time-consuming, and uncertain.33 Thereafter, to gain marketing approval and satisfy GFATM quality standards, generic companies have to prove bioequivalence at an estimated cost of US$15 million32 or more, and face many regulatory barriers,3638 including: overcoming data exclusivity;39 navigating complicated and expensive registration procedures; selecting a local registration agent;27 facing regulatory incapacity, delay, and corruption;3638 and labelling pursuant to dierent regulatory and language needs.27,40 Because generic companies need to register in every country of sale, registration-related costs are multiplied.41 Even after registration, generic producers must secure a viable local distribution system for their product in every country of sale, because many generic companies do not have in-country distribution channels in many countries, and to develop them is expensive and time-consuming.27 Will the Global Fund maintain its promise as a predictable, large-scale purchaser of medicines? Uncertainty over future payments and their timing can deter generic entry and price.27 Acceptance of donations of just one drug might create uncertainty about whether donations of other drugs will be forthcoming, possibly deterring entry of generic drugs for AIDS, tuberculosis, and malaria in general. One of the great advantages of GFATM, from a generic companys perspective, is that it oers substantial buying power, and long-term and predictable guarantees of payment. If GFATM market clout is undermined because of large-scale donations, the market will look insecure for
www.thelancet.com Vol 373 April 4, 2009

generic forecasters. Accordingly, eects of donations can have ripple consequences that will be amplied at the scale of programming supported by the GFATM. The eect on the generic market for medicines goes beyond the market in a particular country, beyond the particular donated drug, and even beyond the GFATM share of the market. The aggregation of large-scale purchasing power from GFATM and other large initiatives, such as Presidents Emergency Plan for AIDS Relief, has been responsible for pushing strong competition between dierent generic producers. Moreover, the security of sustained funding has motivated rapid entry by generic companies for the production of medicines for priority diseases. All these forecasting uncertainties become worse if innovators are donating products. The donated products inevitably undercut the generic price. Similarly, they necessarily shrink the residual market, at least in the short term. Although in-kind drug donations do not directly aect sunk costs and regulatory barriers, these costs and barriers might seem disproportionate in donation-aected markets.32 Anxiety about the duration or expanded product line of drug donations is the coup de grace to entry of generic drugs. In addition to restricting donations so as to incentivise generic entry, countries should be moving towards durable, ecacious, and safe treatment protocols such as those being recommended by WHO.42,43 However, being captured by the lure of donations is a risk because most people and most governments nd it hard to turn down free goods. Moreover, in countries where many people with AIDS are still waiting for treatment, governments might face strong political pressure, even from treatment activists, to accept free products that could increase numbers of people treated. Although little direct evidence exists of donations eect on government therapeutic choices, free drugs at the scale of GFATM-funded programmes might detract from the development of coherent national drug policies6 or, even worse, result in overcommitment by governments to therapeutically worse treatments or to selection of branded versus generic supplies of medicines. Free products can deter or delay modication of national treatment guidelines to specify therapeutically better products, and can subtly undermine government motivation to weigh best therapeutic options independently. Donations of individually formulated products could deter government use of therapeutically improved xed-dose combinations, many of which are only available in generic form, and of improved formulations of drugs (eg, drugs with improved stability or tolerability). They could also cause premature start of second-line regimens or reduce options for preferred, simplied second-line regimens. Finally, donations of tied products could deter use of therapeutic alternatives. For example, a therapeutic preference or true cost advantage of atazanavir and ritonavir might exist over
1219

Viewpoint

lopinavir and ritonavir, or of tenofovir and lamivudine over tenofovir and emtribicatine, that could be thwarted by donations. Moreover, free products can result in product familiarity, brand loyalty, and previous-use preferences by physicians and patients, and can negatively aect prescriber practices with respect to generic equivalents.7 Donations often bypass drug regulatory processes, creating advantages over generic drugs. They can also cause governments to question the propriety of registering generic equivalents,4 and can distort drug procurement and supply systems where they result in separate arrangements for antiretroviral drugs.4 Even when major pharmaceutical companies have a genuine desire for charitable donations and to act with corporate responsibility, other reasons for donations that mainly serve self-interests might exist. Tactical donations can be used to: (1) shrink market size and the certainty of sizeable GFATM nanced purchases, and essentially under-cut generic drugs; (2) enhance or regain corporate goodwill through favourable publicity; (3) build product loyalty with both patients and clinicians (thereafter, when donations end, prescription habits and patient preferences continue); (4) develop or reinforce drug distribution systems for the benet of their other drugs; (5) obtain tax advantages; (6) retain monopoly access to the private sector;4 and (7) discourage use of trade-related aspects of intellectual property rights (TRIPS) safeguards by both generic producers and developing countries.33 Even more sinister motivations could result in predatory donations timed after generic companies have made substantial investments, crossed regulatory barriers, and secured distribution chains, thus punishing early entrants and warning all other prospective entrants. Our main concern with drug donations is that emerging and more competitive pharmaceutical markets will be distorted by donations that reduce aggregate market size for competing generic products and thus deter entry of generic drugs.26 Donations that produce smaller markets and fewer generic entrants will retard the development of competitive markets in which multiple generic entrants produce at ecient economies of scale, and thereafter compete for sales at or near marginal costs of production. These shrunken-market eects will negatively aect both domestic and worldwide generic producers.7,30 Although entry decisions of major generic producers, such as those in India, will be aected by the aggregate worldwide market,27 negative eects of donations on small local producers, such as those emerging in Africa,4448 will probably be even greater. Moreover, even one donation can create uncertainty about the size of the future generic market for other medicines where donations might happen, and thus donations of a subset of medicines can create uncertainty about worldwide market for an entire priority disease market, especially for antiretroviral drugs.
1220

Treatment activists, multilateral institutions, such as WHO, developing countries, and board members should work together to resist drug donations to GFATM. Failure to defeat this revived donation initiative could pose substantial risks to GFATMs reputation. Additionally, accepting donations could have important negative eects on future costs of life-saving medicines and on future therapeutic choices.
Conict of interest statement BKB is a member of the Contact Group of the Northern non-governmental organisation Delegation to the Board of the GFATM, and is Chairman of Health Global Access Project, an HIV/AIDS activist organisation. EO is the Coordinator of the Ecumenical Pharmaceutical Network and an adviser of World Council of Churches. We declare that we have no conict of interest. References 1 The Global Fund to Fight AIDS. Tuberculosis and Malaria. Summary of Global Fund Board discussions and decisions related to product/service donation. 2007. 2 World Health Organization. Guidelines for Drug Donations. Revised 1999. http://whqlibdoc.who.int/hq/1999/WHO_EDM_ PAR_99.4.pdf (accessed Jan 25, 2008). 3 World Health Organization. First-year experiences with the Interagency Guidelines for Drug Donations. 2000. http://whqlibdoc.who.int/hq/2000/WHO_EDM_PAR_2000.1.pdf (accessed Jan 25, 2008). 4 Caines K, Lush L. Impact of publicprivate partnerships addressing access to pharmaceuticals in selected low and middle income countries: a synthesis report from studies in Botswana, Sri Lanka, Uganda and Zambia. Initiative on PublicPrivate Partnerships for Health. 2004. http://www.dd.gov.uk/Pubs/les/ ipph-accesspharmaceuticals-synthesis.pdf (accessed Jan 25, 2008). 5 Caines K. Global health partnerships and neglected diseases, DfID Health Resource Center. 2004. http://www2.ohchr.org/ english/issues/development/docs/WHO_4.pdf (accessed Jan 25, 2008). 6 Shretta R, Walt G, Brugha R, Snow RW. A political analysis of corporate drug donations: the example of Malarone in Kenya. Health Policy Plan 2001; 16: 16170. 7 Hechmann R, Bunde-Birouste A. Drug donations in emergencies, the Sri Lankan post-tsunami experience. J Humanitarian Asst 2007. http://jha.ac/2007/09/26/drug-donations-in-emergencies-the-srilankan-post-tsunami-experience (accessed Jan 25, 2008). 8 Autier P, Govindarah R, Gray R, Lakshminarayanan R, Nassery HG, Schmets G. Drug donations in post-emergency situations. World Bank 2002. http://siteresources.worldbank.org/HEALTHNUTRITI ONANDPOPULATION/Resources/281627-1095698140167/NasseryDrugDonation-whole.pdf (accessed Jan 25, 2008). 9 Guilloux A, Moon S. Hidden price tags: disease-specic drug donations: costs and alternatives. 2001. 10 World Health Organization. Global strategy for containment of antimicrobial resistance. 2001. http://www.who.int/csr/resources/ publications/drugresist/en/EGlobal_Strat.pdf (accessed Jan 25, 2008). 11 Oyediran ABO, Ddumba E, Ochola, S, Lucas AO, Koporc K, Dowdle WR. A publicprivate partnership for malaria control: lessons from the Malarone Donation Programme. Bull World Health Organ 2002; 80: 81721. 12 Congressional Budget Oce. How increased competition from generic drugs has aected prices and returns in the pharmaceutical industry. Congress of the United States, July, 1998. http://www.cbo. gov/ftpdocs/6xx/doc655/pharm.pdf (accessed Jan 29, 2008). 13 Reien D, Ward MR. Generic drug industry dynamics. Rev Econ Stat 2005; 87: 3749. 14 Bae J. Research on pharmaceutical drug development, use, and outcomes: drug patent expirations and the speed of generic entry. Health Services Res 1997; 32: 87101. 15 Dalen DM, Strm S, Haabeth T. Price regulation and generic competition in the pharmaceutical market. University of Oslo Health Econ Res Prog 2006. http://www.hero.uio.no/publicat/2006/ HERO2006_1.pdf (accessed Jan 29, 2008).

www.thelancet.com Vol 373 April 4, 2009

Viewpoint

16 17

18

19

20 21 22

23 24

25 26

27

28

29

30

31

32

33

Morton F. Entry decisions in the generic pharmaceutical industry. RAND J Econ 1999; 30: 42140. Hass J, Phillips KA, Gerstenberger E, Seger A. Potential savings from substituting generic drugs for brand-name drugs: medical expenditure panel survey, 19972000. Ann Intern Med 2005; 142: 89197. Saha A, Grabowski H, Birnbaum H, Greenberg P, Bizan O. Generic competition in the US pharmaceutical industry. Int J Econ Bus 2006; 13: 1538. Reien D, Ward M. Branded generics as a strategy to limit cannibalization of pharmaceutical markets. 2005. http://www.uta. edu/faculty/mikeward/brandedgenerics.pdf (accessed Jan 29, 2008). Frank RG. The ongoing regulation of generic drugs. N Engl J Med 2007; 357: 199396. Henry D, Lexchin J. The pharmaceutical industry as a medicines provider. Lancet 2002; 360: 159094. Moreno-Torres I, Puig-Junoy J, Borrell-Arqu J-R. Generic entry into a regulated pharmaceutical market. 2007. http://ssrn.colm/ abstract=992685 (accessed Jan 25, 2008). Frank RG, Salkever DS. Generic entry and pricing of pharmaceuticals. J Econ Manage Strategy 1997; 6: 7590. Hudson J. Generic take-up in the pharmaceutical market following patent expiry: a multi-country study. Int Rev Law Econ 2000; 24: 10312. Rudholm N. Entry and the number of rms in the Swedish pharmaceutical market. Rev Ind Organ 2001; 19: 35164. Lewis-Lettington R, Munyi P. Willingness and ability to use TRIPS exibilities: Kenya Case Study. DfID, Sept 2004. http://www.dd. gov.uk/pubs/les/ddkenyareport.pdf (accessed Jan 25, 2008). World Community Advisory Board. Meeting with Indian generic drugs industry. Jan 67, 2005. http://www.i-base.info/wcab/PDF/ W-CAB-Mumbai-report-jul05.pdf (accessed Jan 25, 2008). Chien C. HIV/AIDS drugs for sub-Saharan Africa: how do brand and generic supply compare? PLoS One, March 2007. http://www. pubmedcentral.nih.gov/picrender.fcgi?artid=1805689&blobtype=pdf (accessed Jan 29, 2008). Mdecins Sans Frontires. Untangling the web of price reductions: a pricing guide for the purchase of ARVs for developing countries. Revised Sept 2007. http://www.accessmed-msf.org/leadmin/user_ upload/diseases/hiv-aids/UTW10_RSep_horizontal.pdf (accessed Jan 29, 2008). World Health Organization. A summary report from the global price reporting mechanism on antiretroviral drugs. July 2007. http://www.who.int/hiv/amds/GPRMsummaryJuly2007.pdf (accessed Jan 29, 2008). Prez-Casas C, Herranz E, Ford N. Pricing of drugs and donations: options for sustainable equity pricing. Trop Med Int Health 2001; 6: 96064. Competition Bureau Canada. Canadian generic drug sector study. Oct 2007. http://www.competitionbureau.gc.ca/epic/site/cb-bc.nsf/ vwapj/Competition%20Bureau%20Generic%20Drug%20Sector%20 Study.pdf/$FILE/Competition%20Bureau%20Generic%20Drug%20 Sector%20Study.pdf (accessed Feb 1, 2008). Baker B. Processes and issues for improving access to medicines: willingness and ability to use TRIPS-compliant exibilities in developing countries. DfID Aug 2004. http://www.dd.gov.uk/ pubs/les/ddbrookbakertrips.pdf (accessed Feb 1, 2008).

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

Thorpe P. Study on the implementation of the TRIPS Agreement in developing countries. 2001. http://www.iprcommission.org/ papers/pdfs/study_papers/sp7_thorpe_study.pdf (accessed Dec 14, 2007). Musungu S, Oh C. The use of exibilities in TRIPS by developing countries: can they promote access to medicines? 2006. http://www. southcentre.org/publications/SouthPerspectiveSeries/TheUseOfFle xibilitiesInTripsFinal.pdf (accessed Dec 14, 2007). World Health Organization. Measuring transparency in medicines registration, selection and procurement: four country assessment studies. 2006. http://www.who.int/medicines/areas/policy/ goodgovernance/Transparency4CountryStudy.pdf (accessed Dec 14, 2007). Ratanawitrasin S, Wondemagegnehu E. Eective drug registration: a multi-country study. 2002. http://whqlibdoc.who.int/ hq/2002/9241562064.pdf (accessed Dec 14, 2007). European Community. Better regulation of pharmaceuticals: towards a simpler, clearer and more exible framework on variations. Oct 20, 2006. http://ec.europa.eu/enterprise/ pharmaceuticals/varreg/variations_consultation_paper_codecision_ 20070710.pdf (accessed Dec 14, 2007). Correa C. Protection of data submitted for the registration of pharmaceuticals: implementing the standards of the TRIPS Agreement. 2002. Kimani D. Drug companies ght over Kenyan market. Nationmedia.com Jan 23, 2006. http://www.nationmedia.com/ eastafrican/23012006/News/news230120067.htm (accessed Feb 4, 2008). Thompson MB. Proceedings: HIV/AIDS donations and price reductions in sub-Saharan Africa. Global Alliance for Womens Health April 15, 2003. http://www.gawh.org/programmes/04_15_ 03/pdf/proceedings.pdf?CAMEFROM=SYMPOSIA&id=13 (accessed Jan 25, 2008). World Health Organization. Prioritizing second-line antiretroviral drugs for adults and adolescents: a public health approach 2007. http://www.who.int/hiv/pub/meetingreports/second_line_art_ report_2008.pdf (accessed Feb 4, 2008). World Health Organization. Antiretroviral therapy for HIV infection in adults and adolescents: a public health approach 2007. http://www.who.int/hiv/pub/guidelines/artadultguidelines.pdf (accessed Feb 4, 2008). Uganda: Factory to boost ARV rollout. Plus News, Oct 10, 2007. http://www.plusnews.org/Report.aspx?ReportId=74715 (accessed Feb 4, 2008). Mangwiro C. Brazil oers drug factory to AIDS ravished Mozambique. Reuters AlertNet May 29, 2007. http://www.alertnet. org/thenews/newsdesk/L29579337.htm (accessed Feb 4, 2008). Kahn T. Ranbaxy, CIH in sub-Saharan antiretroviral joint venture. Bus Day (Johannesburg) Feb 7, 2006. http://www.businessday.co.za/ articles/article.aspx?ID=BD4A150867 (accessed Feb 4, 2008). Kahn T. Adcock Ingram mounts AIDS challenge. Bus Day (Johannesburg) Feb 7, 2006. http://www.businessday.co.za/articles/ economy.aspx?ID=BD4A396084 (accessed Feb 4, 2008). Kimani D. Drug companies ght over Kenyan market. Nationmedia.com Jan 23, 2006. http://www.nationmedia.com/ eastafrican/23012006/News/news230120067.htm (accessed Feb 4, 2008).

www.thelancet.com Vol 373 April 4, 2009

1221

Case Report

Attentiongrapefruit!
Lucinda A Grande, Raul D Mendez, Richard T Krug, Evert-Jan Verschuyl
Lancet 2009; 373: 1222 St Peter Family Medicine (L A Grande MD), Department of Medicine (R D Mendez MD), Department of Surgery (R T Krug MD), and Department of Interventional Radiology (E-J Verschuyl MD), Providence St Peter Hospital, Olympia, WA 98506 , USA Correspondence to: Dr Lucinda A Grande, St Peter Family Medicine, 525 Lilly Road NE, Olympia, WA 98506, USA lucinda.grande@providence.org

In November, 2008, a 42-year-old woman presented to our emergency department with diculty in walking, shortness of breath, and light-headedness. The day before, she had gone on a 15 h car journey, after which she felt pain in her lower back and left buttock radiating down to the ankle. The following morning, her left leg had turned purple. Medical history was unremarkable. Her medications included levothyroxine, which she had taken for 4 years, and a low-dose combined oral contraceptive containing drospirenone and ethinylestradiol, which she had taken for 1 year. She did not smoke and rarely drank alcohol. Slightly overweight, she had begun an aggressive weight-loss diet 3 days earlier, including 225 g grapefruit every morning; previously she had eaten grapefruit rarely. In the emergency department, her left leg was diusely rm and swollen, with bluish discoloration and livedo reticularis throughout. There were palpable femoral, popliteal, and dorsalis pedis pulses. Duplex ultrasonography showed a deep venous thrombosis, which extended from the external iliac vein distal to the calf veins. Because of the presence of phlegmasia caerulea dolens, there was concern for potential irreversible venous gangrene and subsequent limb loss; an intravenous infusion of unfractionated heparin was immediately started, and the patient underwent catheterdirected thrombolysis with recombinant tissue plasminogen activator. On contrast radiography, stenosis was evident within the left common iliac vein at approximately the 5th lumbar vertebra (gure), suggestive of May-Thurner syndrome. A stent was placed across the stenosis. Leg pain and discoloration completely resolved by that evening; venography 24 h later conrmed a patent vein and stent. She was discharged home on bridging anticoagulation, after beginning a 6-month course of
B

warfarin; she also stopped using her oral contraceptive. A thrombophilia screen done in the outpatient setting was positive for factor V Leiden mutation. When last seen in February, 2009, she was asymptomatic. The May-Thurner syndrome was described in 1957,1 but Virchow rst hypothesised such a disorder in 1851 on the basis that the risk of deep venous thrombosis is more than three times higher in the left than in the right leg.2 The syndrome results from compression of the left common iliac vein between the overlying right common iliac artery and the underlying 5th lumbar vertebral body. Chronic pulsation of the artery results in intimal hyperplasia within the vein. Iliofemoral vein thrombosis eventually occurs. Left iliac vein stenosis is not rare, but for thrombosis to occur, all three components of Virchows triad must be present: stasis, endothelial injury, and hypercoagulability. Procoagulant contributors in our patient included factor V Leiden mutation and oestrogen contraceptive use.3 What tipped the balance toward acute venous thrombosis that day? It is likely that stasis occurred as our patients already stenosed iliac vein was compressed further by hip exion during her car journey. We hypothesise that she also had enhanced hypercoagulability as a consequence of her new diet. Grapefruit juice can augment the bioavailability of ethinylestradiol4 by inhibiting CYP3A4, a cytochrome P450 enzyme densely expressed in the wall of the small bowel. The enzyme metabolises many drugs, including ethinylestradiol, dihydropyridines, and statins. Inhibition of CYP3A4 by grapefruit juice most likely occurs through increased enzyme degradation. With an eective half-life of about 12 h, a cumulative eect of daily grapefruit juice consumption can occur. The magnitude of eect varies; those with high inherent bowel wall enzyme content show the largest enzyme reduction with juice treatment, resulting in an increase in peak drug concentration of up to six times.4 Our patient had a constellation of potential risk factors for venous thrombosis; a heightened hypercoagulable state from increased ethinylestradiol serum concentration due to her 3 days of grapefruit for breakfast may well have tipped the balance.
Contributors LAG prepared the report. RTK evaluated the patient in the emergency room. E-J V performed the catheterisation procedure; LAG and RDM managed the patient in the hospital after the procedure. References 1 May R, Thurner J. The cause of predominantly sinistral occurrence of thrombosis of the pelvic veins. Angiology 1957; 8: 41927. 2 Virchow R. Uber die Erweiterung kleiner Gefasse. Arch Pathol Anat 1851; 3: 427. 3 Vandenbroucke JP, Rosing J, Bloemenkamp KW, et al. Oral contraceptives and the risk of venous thrombosis. N Engl J Med 2001; 344: 152735. 4 Bailey DG, Arnold MO, Spence JD. Grapefruit juice-drug interactions. Br J Clin Pharmacol 1998; 46: 10110.

Figure: Contrast radiography showing stenosis within the left common iliac vein (A) The rough wall of the stenosed vessel is evidence of endothelial injury. (B) The stenosis is located at the 5th lumbar vertebra; this is the same view as (A) but with dierent contrast settings.

1222

www.thelancet.com Vol 373 April 4, 2009